repo_id
stringlengths 21
96
| file_path
stringlengths 31
155
| content
stringlengths 1
92.9M
| __index_level_0__
int64 0
0
|
|---|---|---|---|
rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm | rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm/mlflow-tracking-server/values.yaml | # Default values for mlflow-tracking-server.
# This is a YAML-formatted file.
# Declare variables to be passed into your templates.
replicaCount: 1
env:
mlflowArtifactPath: ""
mlflowUser: "postgres"
mlflowPass: "mlflow"
mlflowDBName: "mlflow_db"
mlflowDBAddr: "mlf-db-postgresql"
mlflowDBPort: "5432"
image:
repository: ""
pullPolicy: IfNotPresent
tag: "gcp"
imagePullSecrets: []
nameOverride: ""
fullnameOverride: ""
serviceAccount:
create: false
annotations: {}
name: ""
podAnnotations: {}
podSecurityContext: {}
securityContext: {}
service:
type: ClusterIP
port: 80
ingress:
enabled: false
annotations: {}
hosts:
- host: chart-example.local
paths: []
tls: []
resources: {}
autoscaling:
enabled: false
minReplicas: 1
maxReplicas: 100
targetCPUUtilizationPercentage: 80
nodeSelector: {}
tolerations: []
affinity: {}
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm/mlflow-tracking-server | rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm/mlflow-tracking-server/templates/_helpers.tpl | {{/*
Expand the name of the chart.
*/}}
{{- define "mlflow-tracking-server.name" -}}
{{- default .Chart.Name .Values.nameOverride | trunc 63 | trimSuffix "-" }}
{{- end }}
{{/*
Create a default fully qualified app name.
We truncate at 63 chars because some Kubernetes name fields are limited to this (by the DNS naming spec).
If release name contains chart name it will be used as a full name.
*/}}
{{- define "mlflow-tracking-server.fullname" -}}
{{- if .Values.fullnameOverride }}
{{- .Values.fullnameOverride | trunc 63 | trimSuffix "-" }}
{{- else }}
{{- $name := default .Chart.Name .Values.nameOverride }}
{{- if contains $name .Release.Name }}
{{- .Release.Name | trunc 63 | trimSuffix "-" }}
{{- else }}
{{- printf "%s-%s" .Release.Name $name | trunc 63 | trimSuffix "-" }}
{{- end }}
{{- end }}
{{- end }}
{{/*
Create chart name and version as used by the chart label.
*/}}
{{- define "mlflow-tracking-server.chart" -}}
{{- printf "%s-%s" .Chart.Name .Chart.Version | replace "+" "_" | trunc 63 | trimSuffix "-" }}
{{- end }}
{{/*
Common labels
*/}}
{{- define "mlflow-tracking-server.labels" -}}
helm.sh/chart: {{ include "mlflow-tracking-server.chart" . }}
{{ include "mlflow-tracking-server.selectorLabels" . }}
{{- if .Chart.AppVersion }}
app.kubernetes.io/version: {{ .Chart.AppVersion | quote }}
{{- end }}
app.kubernetes.io/managed-by: {{ .Release.Service }}
{{- end }}
{{/*
Selector labels
*/}}
{{- define "mlflow-tracking-server.selectorLabels" -}}
app.kubernetes.io/name: {{ include "mlflow-tracking-server.name" . }}
app.kubernetes.io/instance: {{ .Release.Name }}
{{- end }}
{{/*
Create the name of the service account to use
*/}}
{{- define "mlflow-tracking-server.serviceAccountName" -}}
{{- if .Values.serviceAccount.create }}
{{- default (include "mlflow-tracking-server.fullname" .) .Values.serviceAccount.name }}
{{- else }}
{{- default "default" .Values.serviceAccount.name }}
{{- end }}
{{- end }}
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm/mlflow-tracking-server | rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm/mlflow-tracking-server/templates/service.yaml | apiVersion: v1
kind: Service
metadata:
name: {{ include "mlflow-tracking-server.fullname" . }}
labels:
{{- include "mlflow-tracking-server.labels" . | nindent 4 }}
spec:
type: {{ .Values.service.type }}
ports:
- port: {{ .Values.service.port }}
targetPort: http
protocol: TCP
name: http
selector:
{{- include "mlflow-tracking-server.selectorLabels" . | nindent 4 }}
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm/mlflow-tracking-server | rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm/mlflow-tracking-server/templates/hpa.yaml | {{- if .Values.autoscaling.enabled }}
apiVersion: autoscaling/v2beta1
kind: HorizontalPodAutoscaler
metadata:
name: {{ include "mlflow-tracking-server.fullname" . }}
labels:
{{- include "mlflow-tracking-server.labels" . | nindent 4 }}
spec:
scaleTargetRef:
apiVersion: apps/v1
kind: Deployment
name: {{ include "mlflow-tracking-server.fullname" . }}
minReplicas: {{ .Values.autoscaling.minReplicas }}
maxReplicas: {{ .Values.autoscaling.maxReplicas }}
metrics:
{{- if .Values.autoscaling.targetCPUUtilizationPercentage }}
- type: Resource
resource:
name: cpu
targetAverageUtilization: {{ .Values.autoscaling.targetCPUUtilizationPercentage }}
{{- end }}
{{- if .Values.autoscaling.targetMemoryUtilizationPercentage }}
- type: Resource
resource:
name: memory
targetAverageUtilization: {{ .Values.autoscaling.targetMemoryUtilizationPercentage }}
{{- end }}
{{- end }}
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm/mlflow-tracking-server | rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm/mlflow-tracking-server/templates/NOTES.txt | 1. Get the application URL by running these commands:
{{- if .Values.ingress.enabled }}
{{- range $host := .Values.ingress.hosts }}
{{- range .paths }}
http{{ if $.Values.ingress.tls }}s{{ end }}://{{ $host.host }}{{ . }}
{{- end }}
{{- end }}
{{- else if contains "NodePort" .Values.service.type }}
export NODE_PORT=$(kubectl get --namespace {{ .Release.Namespace }} -o jsonpath="{.spec.ports[0].nodePort}" services {{ include "mlflow-tracking-server.fullname" . }})
export NODE_IP=$(kubectl get nodes --namespace {{ .Release.Namespace }} -o jsonpath="{.items[0].status.addresses[0].address}")
echo http://$NODE_IP:$NODE_PORT
{{- else if contains "LoadBalancer" .Values.service.type }}
NOTE: It may take a few minutes for the LoadBalancer IP to be available.
You can watch the status of by running 'kubectl get --namespace {{ .Release.Namespace }} svc -w {{ include "mlflow-tracking-server.fullname" . }}'
export SERVICE_IP=$(kubectl get svc --namespace {{ .Release.Namespace }} {{ include "mlflow-tracking-server.fullname" . }} --template "{{"{{ range (index .status.loadBalancer.ingress 0) }}{{.}}{{ end }}"}}")
echo http://$SERVICE_IP:{{ .Values.service.port }}
{{- else if contains "ClusterIP" .Values.service.type }}
export POD_NAME=$(kubectl get pods --namespace {{ .Release.Namespace }} -l "app.kubernetes.io/name={{ include "mlflow-tracking-server.name" . }},app.kubernetes.io/instance={{ .Release.Name }}" -o jsonpath="{.items[0].metadata.name}")
echo "Visit http://127.0.0.1:8080 to use your application"
kubectl --namespace {{ .Release.Namespace }} port-forward $POD_NAME 8080:80
{{- end }}
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm/mlflow-tracking-server | rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm/mlflow-tracking-server/templates/ingress.yaml | {{- if .Values.ingress.enabled -}}
{{- $fullName := include "mlflow-tracking-server.fullname" . -}}
{{- $svcPort := .Values.service.port -}}
{{- if semverCompare ">=1.14-0" .Capabilities.KubeVersion.GitVersion -}}
apiVersion: networking.k8s.io/v1beta1
{{- else -}}
apiVersion: extensions/v1beta1
{{- end }}
kind: Ingress
metadata:
name: {{ $fullName }}
labels:
{{- include "mlflow-tracking-server.labels" . | nindent 4 }}
{{- with .Values.ingress.annotations }}
annotations:
{{- toYaml . | nindent 4 }}
{{- end }}
spec:
{{- if .Values.ingress.tls }}
tls:
{{- range .Values.ingress.tls }}
- hosts:
{{- range .hosts }}
- {{ . | quote }}
{{- end }}
secretName: {{ .secretName }}
{{- end }}
{{- end }}
rules:
{{- range .Values.ingress.hosts }}
- host: {{ .host | quote }}
http:
paths:
{{- range .paths }}
- path: {{ . }}
backend:
serviceName: {{ $fullName }}
servicePort: {{ $svcPort }}
{{- end }}
{{- end }}
{{- end }}
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm/mlflow-tracking-server | rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm/mlflow-tracking-server/templates/deployment.yaml | apiVersion: apps/v1
kind: Deployment
metadata:
name: {{ include "mlflow-tracking-server.fullname" . }}
labels:
{{- include "mlflow-tracking-server.labels" . | nindent 4 }}
spec:
{{- if not .Values.autoscaling.enabled }}
replicas: {{ .Values.replicaCount }}
{{- end }}
selector:
matchLabels:
{{- include "mlflow-tracking-server.selectorLabels" . | nindent 6 }}
template:
metadata:
{{- with .Values.podAnnotations }}
annotations:
{{- toYaml . | nindent 8 }}
{{- end }}
labels:
{{- include "mlflow-tracking-server.selectorLabels" . | nindent 8 }}
spec:
{{- with .Values.imagePullSecrets }}
imagePullSecrets:
{{- toYaml . | nindent 8 }}
{{- end }}
serviceAccountName: {{ include "mlflow-tracking-server.serviceAccountName" . }}
securityContext:
{{- toYaml .Values.podSecurityContext | nindent 8 }}
volumes:
- name: gcsfs-creds
secret:
secretName: gcsfs-creds
items:
- key: keyfile.json
path: keyfile.json
containers:
- name: {{ .Chart.Name }}
securityContext:
{{- toYaml .Values.securityContext | nindent 12 }}
volumeMounts:
- name: gcsfs-creds
mountPath: "/etc/secrets"
readOnly: true
image: "{{ .Values.image.repository }}:{{ .Values.image.tag | default .Chart.AppVersion }}"
imagePullPolicy: {{ .Values.image.pullPolicy }}
ports:
- name: http
containerPort: 80
protocol: TCP
resources:
{{- toYaml .Values.resources | nindent 12 }}
env:
- name: MLFLOW_USER
value: "{{ .Values.env.mlflowUser }}"
- name: MLFLOW_PASS
value: "{{ .Values.env.mlflowPass }}"
- name: MLFLOW_DB_NAME
value: "{{ .Values.env.mlflowDBName }}"
- name: MLFLOW_DB_ADDR
value: "{{ .Values.env.mlflowDBAddr }}"
- name: MLFLOW_DB_PORT
value: "{{ .Values.env.mlflowDBPort }}"
- name: MLFLOW_ARTIFACT_PATH
value: "{{ .Values.env.mlflowArtifactPath }}"
- name: GOOGLE_APPLICATION_CREDENTIALS
value: "/etc/secrets/keyfile.json"
{{- with .Values.nodeSelector }}
nodeSelector:
{{- toYaml . | nindent 8 }}
{{- end }}
{{- with .Values.affinity }}
affinity:
{{- toYaml . | nindent 8 }}
{{- end }}
{{- with .Values.tolerations }}
tolerations:
{{- toYaml . | nindent 8 }}
{{- end }}
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm/mlflow-tracking-server | rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/helm/mlflow-tracking-server/templates/serviceaccount.yaml | {{- if .Values.serviceAccount.create -}}
apiVersion: v1
kind: ServiceAccount
metadata:
name: {{ include "mlflow-tracking-server.serviceAccountName" . }}
labels:
{{- include "mlflow-tracking-server.labels" . | nindent 4 }}
{{- with .Values.serviceAccount.annotations }}
annotations:
{{- toYaml . | nindent 4 }}
{{- end }}
{{- end }}
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment | rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/images/network_diagram.svg | <svg version="1.1" viewBox="0.0 0.0 536.1338582677165 391.21522309711287" fill="none" stroke="none" stroke-linecap="square" stroke-miterlimit="10" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns="http://www.w3.org/2000/svg"><clipPath id="p.0"><path d="m0 0l536.13385 0l0 391.2152l-536.13385 0l0 -391.2152z" clip-rule="nonzero"/></clipPath><g clip-path="url(#p.0)"><path fill="#000000" fill-opacity="0.0" d="m0 0l536.13385 0l0 391.2152l-536.13385 0z" fill-rule="evenodd"/><path fill="#4a86e8" d="m326.33234 302.77716l0 0c-1.2258606 -8.192322 2.7988586 -16.302216 10.366333 -20.888367c7.567505 -4.586151 17.350555 -4.8442383 25.197876 -0.6647644l0 0c2.7797546 -4.763336 7.867523 -8.0520935 13.724365 -8.8714905c5.856842 -0.819397 11.79483 0.92681885 16.017853 4.710449l0 0c2.367981 -4.318756 7.0176086 -7.220398 12.298889 -7.675293c5.2812805 -0.45489502 10.446838 1.6013489 13.663635 5.439087l0 0c4.278137 -4.577881 11.084747 -6.505371 17.47461 -4.948395c6.3898315 1.5569458 11.2152405 6.318695 12.388184 12.2247925l0 0c5.241455 1.3001099 9.607483 4.605255 11.970001 9.061401c2.3625488 4.4561768 2.4898682 9.626221 0.34906006 14.174438l0 0c5.1612854 6.1087036 6.3686523 14.248505 3.1715088 21.381744c-3.1971436 7.1332703 -10.318451 12.188263 -18.70639 13.278595c-0.05911255 6.6947937 -4.0966187 12.837891 -10.556305 16.061401c-6.459656 3.2235413 -14.332794 3.02417 -20.584717 -0.52127075c-2.6629944 8.018188 -10.158447 13.917908 -19.248016 15.150238c-9.089569 1.2323303 -18.143799 -2.423645 -23.250916 -9.388367c-6.260254 3.4329224 -13.772034 4.4218445 -20.84085 2.7436829c-7.068817 -1.6781921 -13.099213 -5.8821106 -16.730835 -11.663452l0 0c-6.397217 0.6807556 -12.582428 -2.3332825 -15.485931 -7.5463257c-2.903534 -5.2130127 -1.9072571 -11.515259 2.4943237 -15.778961l0 0c-5.7064514 -3.0543213 -8.618256 -9.115082 -7.21698 -15.021851c1.4013062 -5.9067993 6.7980957 -10.321075 13.376221 -10.940979z" fill-rule="evenodd"/><path fill="#000000" fill-opacity="0.0" d="m320.04663 329.05573l0 0c2.6929016 1.4413147 5.803833 2.0951538 8.9151 1.8736877m4.0751038 21.452118c1.3379517 -0.14239502 2.649414 -0.44387817 3.9005432 -0.89675903m33.66928 9.816132c-0.94104004 -1.2832947 -1.7289124 -2.6546326 -2.3502197 -4.090637m44.85077 -1.6714478l0 0c0.48550415 -1.461853 0.8000488 -2.9664001 0.93844604 -4.4885254m30.201416 -11.051056c0.062927246 -7.127655 -4.3888245 -13.653809 -11.443115 -16.775269m26.977844 -17.884308c-1.1424255 2.4270935 -2.8864746 4.580139 -5.0953674 6.2903137m-7.2226257 -29.526794l0 0c0.19464111 0.98010254 0.284729 1.9749451 0.26901245 2.9707336m-30.131317 -10.246826l0 0c-1.0671692 1.1419373 -1.9463806 2.4180603 -2.61026 3.7886047m-23.352509 -1.552887l0 0c-0.5687256 1.0372314 -0.9934082 2.1347961 -1.2642517 3.267395m-28.478394 0.8937378l0 0c1.660492 0.8843689 3.1966553 1.9487915 4.5747986 3.1699219m-40.13855 18.3833l0 0c0.16894531 1.1290894 0.43588257 2.2442932 0.79837036 3.3351135" fill-rule="evenodd"/><path stroke="#000000" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m326.33234 302.77716l0 0c-1.2258606 -8.192322 2.7988586 -16.302216 10.366333 -20.888367c7.567505 -4.586151 17.350555 -4.8442383 25.197876 -0.6647644l0 0c2.7797546 -4.763336 7.867523 -8.0520935 13.724365 -8.8714905c5.856842 -0.819397 11.79483 0.92681885 16.017853 4.710449l0 0c2.367981 -4.318756 7.0176086 -7.220398 12.298889 -7.675293c5.2812805 -0.45489502 10.446838 1.6013489 13.663635 5.439087l0 0c4.278137 -4.577881 11.084747 -6.505371 17.47461 -4.948395c6.3898315 1.5569458 11.2152405 6.318695 12.388184 12.2247925l0 0c5.241455 1.3001099 9.607483 4.605255 11.970001 9.061401c2.3625488 4.4561768 2.4898682 9.626221 0.34906006 14.174438l0 0c5.1612854 6.1087036 6.3686523 14.248505 3.1715088 21.381744c-3.1971436 7.1332703 -10.318451 12.188263 -18.70639 13.278595c-0.05911255 6.6947937 -4.0966187 12.837891 -10.556305 16.061401c-6.459656 3.2235413 -14.332794 3.02417 -20.584717 -0.52127075c-2.6629944 8.018188 -10.158447 13.917908 -19.248016 15.150238c-9.089569 1.2323303 -18.143799 -2.423645 -23.250916 -9.388367c-6.260254 3.4329224 -13.772034 4.4218445 -20.84085 2.7436829c-7.068817 -1.6781921 -13.099213 -5.8821106 -16.730835 -11.663452l0 0c-6.397217 0.6807556 -12.582428 -2.3332825 -15.485931 -7.5463257c-2.903534 -5.2130127 -1.9072571 -11.515259 2.4943237 -15.778961l0 0c-5.7064514 -3.0543213 -8.618256 -9.115082 -7.21698 -15.021851c1.4013062 -5.9067993 6.7980957 -10.321075 13.376221 -10.940979z" fill-rule="evenodd"/><path stroke="#000000" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m320.04663 329.05573l0 0c2.6929016 1.4413147 5.803833 2.0951538 8.9151 1.8736877m4.0751038 21.452118c1.3379517 -0.14239502 2.649414 -0.44387817 3.9005432 -0.89675903m33.66928 9.816132c-0.94104004 -1.2832947 -1.7289124 -2.6546326 -2.3502197 -4.090637m44.85077 -1.6714478l0 0c0.48550415 -1.461853 0.8000488 -2.9664001 0.93844604 -4.4885254m30.201416 -11.051056c0.062927246 -7.127655 -4.3888245 -13.653809 -11.443115 -16.775269m26.977844 -17.884308c-1.1424255 2.4270935 -2.8864746 4.580139 -5.0953674 6.2903137m-7.2226257 -29.526794l0 0c0.19464111 0.98010254 0.284729 1.9749451 0.26901245 2.9707336m-30.131317 -10.246826l0 0c-1.0671692 1.1419373 -1.9463806 2.4180603 -2.61026 3.7886047m-23.352509 -1.552887l0 0c-0.5687256 1.0372314 -0.9934082 2.1347961 -1.2642517 3.267395m-28.478394 0.8937378l0 0c1.660492 0.8843689 3.1966553 1.9487915 4.5747986 3.1699219m-40.13855 18.3833l0 0c0.16894531 1.1290894 0.43588257 2.2442932 0.79837036 3.3351135" fill-rule="evenodd"/><path fill="#ffffff" d="m349.06122 319.06226l0 -1.921875l4.984375 0l0 4.5625q-0.734375 0.703125 -2.109375 1.25q-1.375 0.53125 -2.796875 0.53125q-1.796875 0 -3.140625 -0.75q-1.328125 -0.765625 -2.0 -2.171875q-0.671875 -1.40625 -0.671875 -3.046875q0 -1.796875 0.75 -3.1875q0.75 -1.390625 2.203125 -2.125q1.09375 -0.578125 2.734375 -0.578125q2.140625 0 3.328125 0.90625q1.203125 0.890625 1.546875 2.46875l-2.296875 0.421875q-0.234375 -0.84375 -0.90625 -1.328125q-0.671875 -0.484375 -1.671875 -0.484375q-1.515625 0 -2.40625 0.96875q-0.890625 0.953125 -0.890625 2.84375q0 2.03125 0.90625 3.0625q0.90625 1.015625 2.375 1.015625q0.71875 0 1.453125 -0.28125q0.734375 -0.296875 1.25 -0.703125l0 -1.453125l-2.640625 0zm7.1328125 4.21875l0 -11.453125l7.84375 0l0 1.9375l-5.53125 0l0 2.703125l4.78125 0l0 1.9375l-4.78125 0l0 4.875l-2.3125 0zm9.1640625 -3.734375l2.25 -0.21875q0.203125 1.140625 0.828125 1.671875q0.625 0.53125 1.671875 0.53125q1.125 0 1.6875 -0.46875q0.5625 -0.484375 0.5625 -1.109375q0 -0.40625 -0.234375 -0.6875q-0.234375 -0.296875 -0.828125 -0.5q-0.40625 -0.140625 -1.859375 -0.5q-1.859375 -0.46875 -2.609375 -1.140625q-1.046875 -0.9375 -1.046875 -2.296875q0 -0.875 0.484375 -1.640625q0.5 -0.765625 1.4375 -1.15625q0.9375 -0.40625 2.25 -0.40625q2.15625 0 3.25 0.953125q1.09375 0.9375 1.140625 2.515625l-2.3125 0.109375q-0.140625 -0.890625 -0.640625 -1.265625q-0.484375 -0.390625 -1.453125 -0.390625q-1.015625 0 -1.578125 0.40625q-0.375 0.265625 -0.375 0.71875q0 0.40625 0.34375 0.6875q0.4375 0.375 2.125 0.78125q1.6875 0.390625 2.5 0.8125q0.8125 0.421875 1.265625 1.171875q0.453125 0.734375 0.453125 1.8125q0 0.984375 -0.546875 1.84375q-0.546875 0.859375 -1.546875 1.28125q-1.0 0.421875 -2.484375 0.421875q-2.171875 0 -3.34375 -1.0q-1.15625 -1.015625 -1.390625 -2.9375z" fill-rule="nonzero"/><path fill="#000000" fill-opacity="0.0" d="m272.53543 57.637794l0 262.3622" fill-rule="evenodd"/><path stroke="#000000" stroke-width="3.0" stroke-linejoin="round" stroke-linecap="butt" d="m272.53543 57.637794l0 262.3622" fill-rule="evenodd"/><path fill="#c9daf8" d="m359.33032 53.065617l148.09448 0l0 34.488186l-148.09448 0z" fill-rule="evenodd"/><path stroke="#000000" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m359.33032 53.065617l148.09448 0l0 34.488186l-148.09448 0z" fill-rule="evenodd"/><path fill="#ffffff" d="m369.5022 76.16971l0 -11.453125l3.703125 0q2.109375 0 2.75 0.171875q0.984375 0.25 1.640625 1.125q0.671875 0.859375 0.671875 2.21875q0 1.0625 -0.390625 1.78125q-0.375 0.71875 -0.96875 1.125q-0.578125 0.40625 -1.1875 0.546875q-0.828125 0.15625 -2.40625 0.15625l-1.5 0l0 4.328125l-2.3125 0zm2.3125 -9.515625l0 3.25l1.265625 0q1.359375 0 1.8125 -0.171875q0.46875 -0.1875 0.734375 -0.5625q0.265625 -0.390625 0.265625 -0.90625q0 -0.625 -0.375 -1.03125q-0.359375 -0.40625 -0.921875 -0.5q-0.421875 -0.078125 -1.671875 -0.078125l-1.109375 0zm7.828125 5.25q0 -1.09375 0.53125 -2.109375q0.546875 -1.03125 1.53125 -1.5625q1.0 -0.546875 2.21875 -0.546875q1.875 0 3.078125 1.21875q1.203125 1.21875 1.203125 3.09375q0 1.875 -1.21875 3.125q-1.203125 1.234375 -3.046875 1.234375q-1.140625 0 -2.1875 -0.515625q-1.03125 -0.515625 -1.578125 -1.515625q-0.53125 -1.0 -0.53125 -2.421875zm2.25 0.109375q0 1.234375 0.578125 1.890625q0.59375 0.65625 1.453125 0.65625q0.859375 0 1.4375 -0.65625q0.59375 -0.65625 0.59375 -1.90625q0 -1.21875 -0.59375 -1.875q-0.578125 -0.65625 -1.4375 -0.65625q-0.859375 0 -1.453125 0.65625q-0.578125 0.65625 -0.578125 1.890625zm7.2578125 1.78125l2.203125 -0.328125q0.140625 0.640625 0.5625 0.96875q0.4375 0.328125 1.21875 0.328125q0.84375 0 1.28125 -0.3125q0.28125 -0.21875 0.28125 -0.578125q0 -0.25 -0.15625 -0.421875q-0.15625 -0.15625 -0.734375 -0.28125q-2.65625 -0.59375 -3.359375 -1.078125q-0.984375 -0.671875 -0.984375 -1.859375q0 -1.078125 0.84375 -1.8125q0.859375 -0.734375 2.640625 -0.734375q1.703125 0 2.53125 0.5625q0.828125 0.546875 1.140625 1.625l-2.0625 0.390625q-0.140625 -0.484375 -0.515625 -0.734375q-0.359375 -0.265625 -1.046875 -0.265625q-0.875 0 -1.25 0.234375q-0.25 0.171875 -0.25 0.453125q0 0.234375 0.21875 0.390625q0.296875 0.21875 2.046875 0.625q1.765625 0.390625 2.453125 0.96875q0.6875 0.59375 0.6875 1.640625q0 1.140625 -0.953125 1.96875q-0.953125 0.8125 -2.8125 0.8125q-1.703125 0 -2.6875 -0.6875q-0.984375 -0.6875 -1.296875 -1.875zm13.4765625 -5.921875l0 1.75l-1.5 0l0 3.34375q0 1.015625 0.046875 1.1875q0.046875 0.15625 0.1875 0.265625q0.15625 0.109375 0.375 0.109375q0.3125 0 0.890625 -0.203125l0.1875 1.703125q-0.765625 0.328125 -1.734375 0.328125q-0.59375 0 -1.078125 -0.203125q-0.46875 -0.203125 -0.703125 -0.515625q-0.21875 -0.3125 -0.296875 -0.859375q-0.078125 -0.375 -0.078125 -1.546875l0 -3.609375l-1.0 0l0 -1.75l1.0 0l0 -1.65625l2.203125 -1.28125l0 2.9375l1.5 0zm1.328125 8.84375l2.5 0.3125q0.0625 0.4375 0.296875 0.59375q0.3125 0.234375 0.984375 0.234375q0.859375 0 1.28125 -0.25q0.296875 -0.171875 0.4375 -0.5625q0.109375 -0.265625 0.109375 -1.015625l0 -1.203125q-0.984375 1.34375 -2.484375 1.34375q-1.671875 0 -2.65625 -1.421875q-0.765625 -1.109375 -0.765625 -2.78125q0 -2.078125 1.0 -3.171875q1.015625 -1.109375 2.5 -1.109375q1.546875 0 2.546875 1.34375l0 -1.15625l2.046875 0l0 7.4375q0 1.484375 -0.25 2.203125q-0.234375 0.734375 -0.671875 1.140625q-0.4375 0.421875 -1.171875 0.65625q-0.71875 0.234375 -1.84375 0.234375q-2.109375 0 -3.0 -0.734375q-0.875 -0.71875 -0.875 -1.828125q0 -0.109375 0.015625 -0.265625zm1.953125 -4.875q0 1.328125 0.515625 1.9375q0.515625 0.609375 1.265625 0.609375q0.796875 0 1.34375 -0.625q0.5625 -0.625 0.5625 -1.859375q0 -1.296875 -0.53125 -1.921875q-0.53125 -0.625 -1.34375 -0.625q-0.78125 0 -1.296875 0.625q-0.515625 0.609375 -0.515625 1.859375zm10.1171875 4.328125l-2.1875 0l0 -8.296875l2.03125 0l0 1.171875q0.53125 -0.828125 0.9375 -1.09375q0.421875 -0.265625 0.953125 -0.265625q0.75 0 1.453125 0.40625l-0.6875 1.921875q-0.546875 -0.359375 -1.03125 -0.359375q-0.453125 0 -0.78125 0.25q-0.3125 0.25 -0.5 0.921875q-0.1875 0.65625 -0.1875 2.78125l0 2.5625zm8.9296875 -2.640625l2.1875 0.359375q-0.421875 1.203125 -1.328125 1.84375q-0.90625 0.625 -2.28125 0.625q-2.15625 0 -3.203125 -1.421875q-0.8125 -1.125 -0.8125 -2.859375q0 -2.0625 1.078125 -3.21875q1.078125 -1.171875 2.71875 -1.171875q1.859375 0 2.921875 1.21875q1.078125 1.21875 1.03125 3.75l-5.5 0q0.015625 0.96875 0.515625 1.515625q0.515625 0.546875 1.28125 0.546875q0.515625 0 0.859375 -0.28125q0.359375 -0.28125 0.53125 -0.90625zm0.125 -2.21875q-0.015625 -0.953125 -0.484375 -1.453125q-0.46875 -0.5 -1.140625 -0.5q-0.71875 0 -1.1875 0.53125q-0.46875 0.515625 -0.46875 1.421875l3.28125 0zm3.3984375 1.125l2.25 -0.21875q0.203125 1.140625 0.828125 1.671875q0.625 0.53125 1.671875 0.53125q1.125 0 1.6875 -0.46875q0.5625 -0.484375 0.5625 -1.109375q0 -0.40625 -0.234375 -0.6875q-0.234375 -0.296875 -0.828125 -0.5q-0.40625 -0.140625 -1.859375 -0.5q-1.859375 -0.46875 -2.609375 -1.140625q-1.046875 -0.9375 -1.046875 -2.296875q0 -0.875 0.484375 -1.640625q0.5 -0.765625 1.4375 -1.15625q0.9375 -0.40625 2.25 -0.40625q2.15625 0 3.25 0.953125q1.09375 0.9375 1.140625 2.515625l-2.3125 0.109375q-0.140625 -0.890625 -0.640625 -1.265625q-0.484375 -0.390625 -1.453125 -0.390625q-1.015625 0 -1.578125 0.40625q-0.375 0.265625 -0.375 0.71875q0 0.40625 0.34375 0.6875q0.4375 0.375 2.125 0.78125q1.6875 0.390625 2.5 0.8125q0.8125 0.421875 1.265625 1.171875q0.453125 0.734375 0.453125 1.8125q0 0.984375 -0.546875 1.84375q-0.546875 0.859375 -1.546875 1.28125q-1.0 0.421875 -2.484375 0.421875q-2.171875 0 -3.34375 -1.0q-1.15625 -1.015625 -1.390625 -2.9375zm20.484375 2.28125q0.84375 0.609375 1.84375 0.96875l-0.84375 1.640625q-0.53125 -0.15625 -1.03125 -0.4375q-0.109375 -0.046875 -1.53125 -1.015625q-1.125 0.5 -2.5 0.5q-2.640625 0 -4.140625 -1.5625q-1.484375 -1.5625 -1.484375 -4.375q0 -2.796875 1.5 -4.359375q1.5 -1.5625 4.0625 -1.5625q2.546875 0 4.03125 1.5625q1.5 1.5625 1.5 4.359375q0 1.484375 -0.40625 2.609375q-0.3125 0.859375 -1.0 1.671875zm-1.859375 -1.3125q0.4375 -0.515625 0.65625 -1.25q0.234375 -0.75 0.234375 -1.71875q0 -1.984375 -0.875 -2.953125q-0.875 -0.984375 -2.296875 -0.984375q-1.40625 0 -2.296875 0.984375q-0.875 0.984375 -0.875 2.953125q0 2.015625 0.875 3.015625q0.890625 0.984375 2.234375 0.984375q0.5 0 0.9375 -0.15625q-0.703125 -0.46875 -1.421875 -0.71875l0.640625 -1.328125q1.140625 0.390625 2.1875 1.171875zm5.1484375 2.765625l0 -11.359375l2.3125 0l0 9.421875l5.75 0l0 1.9375l-8.0625 0z" fill-rule="nonzero"/><path fill="#4a86e8" d="m359.33032 40.351707l59.370087 0l0 16.598423l-59.370087 0z" fill-rule="evenodd"/><path stroke="#000000" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m359.33032 40.351707l59.370087 0l0 16.598423l-59.370087 0z" fill-rule="evenodd"/><path fill="#ffffff" d="m374.19788 53.73092l0 -10.484375l3.40625 0q1.921875 0 2.515625 0.15625q0.90625 0.234375 1.515625 1.03125q0.609375 0.78125 0.609375 2.03125q0 0.96875 -0.359375 1.625q-0.34375 0.65625 -0.890625 1.03125q-0.53125 0.375 -1.09375 0.5q-0.765625 0.15625 -2.203125 0.15625l-1.375 0l0 3.953125l-2.125 0zm2.125 -8.71875l0 2.984375l1.15625 0q1.25 0 1.671875 -0.15625q0.421875 -0.171875 0.65625 -0.515625q0.25 -0.359375 0.25 -0.828125q0 -0.578125 -0.34375 -0.9375q-0.328125 -0.375 -0.84375 -0.46875q-0.375 -0.078125 -1.53125 -0.078125l-1.015625 0zm7.2287292 3.53125q0 -1.59375 0.46875 -2.6875q0.359375 -0.796875 0.984375 -1.4375q0.625 -0.640625 1.359375 -0.9375q0.984375 -0.421875 2.25 -0.421875q2.328125 0 3.71875 1.4375q1.390625 1.4375 1.390625 4.0q0 2.546875 -1.390625 3.984375q-1.375 1.421875 -3.6875 1.421875q-2.328125 0 -3.71875 -1.421875q-1.375 -1.421875 -1.375 -3.9375zm2.171875 -0.0625q0 1.78125 0.828125 2.703125q0.828125 0.921875 2.09375 0.921875q1.265625 0 2.078125 -0.90625q0.8125 -0.921875 0.8125 -2.75q0 -1.796875 -0.796875 -2.6875q-0.78125 -0.890625 -2.09375 -0.890625q-1.3125 0 -2.125 0.90625q-0.796875 0.890625 -0.796875 2.703125zm9.650085 -5.234375l3.875 0q1.296875 0 1.984375 0.1875q0.921875 0.28125 1.578125 0.984375q0.671875 0.6875 1.015625 1.6875q0.34375 1.0 0.34375 2.484375q0 1.296875 -0.328125 2.234375q-0.390625 1.140625 -1.125 1.84375q-0.546875 0.546875 -1.484375 0.84375q-0.703125 0.21875 -1.875 0.21875l-3.984375 0l0 -10.484375zm2.109375 1.765625l0 6.953125l1.59375 0q0.875 0 1.28125 -0.109375q0.515625 -0.125 0.84375 -0.421875q0.34375 -0.3125 0.5625 -1.015625q0.21875 -0.703125 0.21875 -1.921875q0 -1.21875 -0.21875 -1.859375q-0.21875 -0.65625 -0.609375 -1.015625q-0.375 -0.375 -0.96875 -0.5q-0.453125 -0.109375 -1.75 -0.109375l-0.953125 0z" fill-rule="nonzero"/><path fill="#000000" fill-opacity="0.0" d="m448.39462 1.3385826l83.58267 0l0 86.21526l-83.58267 0z" fill-rule="evenodd"/><g transform="matrix(0.15478267716535432 0.0 0.0 0.15478503937007873 448.39462204724407 1.3385826771653544)"><clipPath id="p.1"><path d="m-4.5474735E-13 0l540.0 0l0 557.0l-540.0 0z" clip-rule="evenodd"/></clipPath><image clip-path="url(#p.1)" fill="#000" width="540.0" height="557.0" x="0.0" y="0.0" preserveAspectRatio="none" xlink:href="data:image/png;base64,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"/></g><path fill="#c9daf8" d="m315.1601 159.65019l148.09451 0l0 34.48819l-148.09451 0z" fill-rule="evenodd"/><path stroke="#000000" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m315.1601 159.65019l148.09451 0l0 34.48819l-148.09451 0z" fill-rule="evenodd"/><path fill="#ffffff" d="m327.6761 182.75429l0 -11.453125l3.453125 0l2.078125 7.8125l2.0625 -7.8125l3.46875 0l0 11.453125l-2.15625 0l0 -9.015625l-2.265625 9.015625l-2.234375 0l-2.265625 -9.015625l0 9.015625l-2.140625 0zm13.421875 0l0 -11.359375l2.3125 0l0 9.421875l5.75 0l0 1.9375l-8.0625 0zm8.7265625 -8.296875l1.21875 0l0 -0.625q0 -1.046875 0.21875 -1.5625q0.234375 -0.515625 0.828125 -0.84375q0.59375 -0.328125 1.515625 -0.328125q0.9375 0 1.828125 0.28125l-0.296875 1.53125q-0.515625 -0.125 -1.0 -0.125q-0.484375 0 -0.6875 0.234375q-0.203125 0.21875 -0.203125 0.84375l0 0.59375l1.640625 0l0 1.71875l-1.640625 0l0 6.578125l-2.203125 0l0 -6.578125l-1.21875 0l0 -1.71875zm6.296875 8.296875l0 -11.453125l2.1875 0l0 11.453125l-2.1875 0zm3.9296875 -4.265625q0 -1.09375 0.53125 -2.109375q0.546875 -1.03125 1.53125 -1.5625q1.0 -0.546875 2.21875 -0.546875q1.875 0 3.078125 1.21875q1.203125 1.21875 1.203125 3.09375q0 1.875 -1.21875 3.125q-1.203125 1.234375 -3.046875 1.234375q-1.140625 0 -2.1875 -0.515625q-1.03125 -0.515625 -1.578125 -1.515625q-0.53125 -1.0 -0.53125 -2.421875zm2.25 0.109375q0 1.234375 0.578125 1.890625q0.59375 0.65625 1.453125 0.65625q0.859375 0 1.4375 -0.65625q0.59375 -0.65625 0.59375 -1.90625q0 -1.21875 -0.59375 -1.875q-0.578125 -0.65625 -1.4375 -0.65625q-0.859375 0 -1.453125 0.65625q-0.578125 0.65625 -0.578125 1.890625zm9.5859375 4.15625l-2.625 -8.296875l2.125 0l1.5625 5.4375l1.421875 -5.4375l2.125 0l1.375 5.4375l1.59375 -5.4375l2.15625 0l-2.65625 8.296875l-2.109375 0l-1.4375 -5.34375l-1.40625 5.34375l-2.125 0zm17.9375 0l0 -9.515625l-3.40625 0l0 -1.9375l9.109375 0l0 1.9375l-3.390625 0l0 9.515625l-2.3125 0zm8.390625 0l-2.1875 0l0 -8.296875l2.03125 0l0 1.171875q0.53125 -0.828125 0.9375 -1.09375q0.421875 -0.265625 0.953125 -0.265625q0.75 0 1.453125 0.40625l-0.6875 1.921875q-0.546875 -0.359375 -1.03125 -0.359375q-0.453125 0 -0.78125 0.25q-0.3125 0.25 -0.5 0.921875q-0.1875 0.65625 -0.1875 2.78125l0 2.5625zm5.7734375 -5.765625l-2.0 -0.359375q0.34375 -1.203125 1.15625 -1.78125q0.828125 -0.578125 2.4375 -0.578125q1.46875 0 2.1875 0.34375q0.71875 0.34375 1.015625 0.890625q0.296875 0.53125 0.296875 1.953125l-0.03125 2.5625q0 1.09375 0.109375 1.625q0.109375 0.515625 0.390625 1.109375l-2.171875 0q-0.078125 -0.21875 -0.203125 -0.65625q-0.0625 -0.1875 -0.078125 -0.25q-0.5625 0.546875 -1.203125 0.828125q-0.640625 0.265625 -1.375 0.265625q-1.28125 0 -2.015625 -0.6875q-0.734375 -0.703125 -0.734375 -1.765625q0 -0.703125 0.328125 -1.25q0.34375 -0.5625 0.9375 -0.84375q0.609375 -0.296875 1.75 -0.515625q1.546875 -0.296875 2.140625 -0.546875l0 -0.21875q0 -0.625 -0.3125 -0.890625q-0.3125 -0.28125 -1.1875 -0.28125q-0.578125 0 -0.90625 0.234375q-0.328125 0.234375 -0.53125 0.8125zm2.9375 1.78125q-0.421875 0.140625 -1.34375 0.34375q-0.90625 0.1875 -1.1875 0.375q-0.4375 0.296875 -0.4375 0.765625q0 0.46875 0.34375 0.8125q0.34375 0.328125 0.875 0.328125q0.59375 0 1.140625 -0.390625q0.390625 -0.296875 0.515625 -0.734375q0.09375 -0.28125 0.09375 -1.0625l0 -0.4375zm11.5546875 -1.859375l-2.171875 0.390625q-0.109375 -0.65625 -0.5 -0.984375q-0.375 -0.328125 -1.0 -0.328125q-0.8125 0 -1.3125 0.578125q-0.484375 0.5625 -0.484375 1.890625q0 1.46875 0.5 2.078125q0.5 0.609375 1.328125 0.609375q0.625 0 1.015625 -0.34375q0.40625 -0.359375 0.578125 -1.234375l2.15625 0.375q-0.34375 1.484375 -1.296875 2.25q-0.953125 0.75 -2.546875 0.75q-1.828125 0 -2.90625 -1.140625q-1.078125 -1.15625 -1.078125 -3.1875q0 -2.0625 1.078125 -3.203125q1.09375 -1.140625 2.9375 -1.140625q1.515625 0 2.40625 0.65625q0.90625 0.640625 1.296875 1.984375zm1.5859375 5.84375l0 -11.453125l2.1875 0l0 6.078125l2.578125 -2.921875l2.703125 0l-2.84375 3.03125l3.046875 5.265625l-2.375 0l-2.078125 -3.734375l-1.03125 1.078125l0 2.65625l-2.1875 0zm8.9765625 -9.421875l0 -2.03125l2.1875 0l0 2.03125l-2.1875 0zm0 9.421875l0 -8.296875l2.1875 0l0 8.296875l-2.1875 0zm11.9921875 0l-2.203125 0l0 -4.234375q0 -1.34375 -0.140625 -1.734375q-0.140625 -0.40625 -0.453125 -0.625q-0.3125 -0.21875 -0.765625 -0.21875q-0.5625 0 -1.015625 0.3125q-0.453125 0.3125 -0.625 0.828125q-0.171875 0.515625 -0.171875 1.90625l0 3.765625l-2.1875 0l0 -8.296875l2.03125 0l0 1.21875q1.09375 -1.40625 2.734375 -1.40625q0.734375 0 1.328125 0.265625q0.609375 0.25 0.90625 0.65625q0.3125 0.40625 0.4375 0.921875q0.125 0.515625 0.125 1.484375l0 5.15625zm2.0234375 0.546875l2.5 0.3125q0.0625 0.4375 0.296875 0.59375q0.3125 0.234375 0.984375 0.234375q0.859375 0 1.28125 -0.25q0.296875 -0.171875 0.4375 -0.5625q0.109375 -0.265625 0.109375 -1.015625l0 -1.203125q-0.984375 1.34375 -2.484375 1.34375q-1.671875 0 -2.65625 -1.421875q-0.765625 -1.109375 -0.765625 -2.78125q0 -2.078125 1.0 -3.171875q1.015625 -1.109375 2.5 -1.109375q1.546875 0 2.546875 1.34375l0 -1.15625l2.046875 0l0 7.4375q0 1.484375 -0.25 2.203125q-0.234375 0.734375 -0.671875 1.140625q-0.4375 0.421875 -1.171875 0.65625q-0.71875 0.234375 -1.84375 0.234375q-2.109375 0 -3.0 -0.734375q-0.875 -0.71875 -0.875 -1.828125q0 -0.109375 0.015625 -0.265625zm1.953125 -4.875q0 1.328125 0.515625 1.9375q0.515625 0.609375 1.265625 0.609375q0.796875 0 1.34375 -0.625q0.5625 -0.625 0.5625 -1.859375q0 -1.296875 -0.53125 -1.921875q-0.53125 -0.625 -1.34375 -0.625q-0.78125 0 -1.296875 0.625q-0.515625 0.609375 -0.515625 1.859375z" fill-rule="nonzero"/><path fill="#4a86e8" d="m315.1601 146.93628l59.370087 0l0 16.59842l-59.370087 0z" fill-rule="evenodd"/><path stroke="#000000" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m315.1601 146.93628l59.370087 0l0 16.59842l-59.370087 0z" fill-rule="evenodd"/><path fill="#ffffff" d="m330.02765 160.31549l0 -10.484375l3.40625 0q1.921875 0 2.515625 0.15625q0.90625 0.234375 1.515625 1.03125q0.609375 0.78125 0.609375 2.03125q0 0.96875 -0.359375 1.625q-0.34375 0.65625 -0.890625 1.03125q-0.53125 0.375 -1.09375 0.5q-0.765625 0.15625 -2.203125 0.15625l-1.375 0l0 3.953125l-2.125 0zm2.125 -8.71875l0 2.984375l1.15625 0q1.25 0 1.671875 -0.15625q0.421875 -0.171875 0.65625 -0.515625q0.25 -0.359375 0.25 -0.828125q0 -0.578125 -0.34375 -0.9375q-0.328125 -0.375 -0.84375 -0.46875q-0.375 -0.078125 -1.53125 -0.078125l-1.015625 0zm7.2287292 3.53125q0 -1.59375 0.46875 -2.6875q0.359375 -0.796875 0.984375 -1.4375q0.625 -0.640625 1.359375 -0.9375q0.984375 -0.421875 2.25 -0.421875q2.328125 0 3.71875 1.4375q1.390625 1.4375 1.390625 4.0q0 2.546875 -1.390625 3.984375q-1.375 1.421875 -3.6875 1.421875q-2.328125 0 -3.71875 -1.421875q-1.375 -1.421875 -1.375 -3.9375zm2.171875 -0.0625q0 1.78125 0.828125 2.703125q0.828125 0.921875 2.09375 0.921875q1.265625 0 2.078125 -0.90625q0.8125 -0.921875 0.8125 -2.75q0 -1.796875 -0.796875 -2.6875q-0.78125 -0.890625 -2.09375 -0.890625q-1.3125 0 -2.125 0.90625q-0.796875 0.890625 -0.796875 2.703125zm9.650116 -5.234375l3.875 0q1.296875 0 1.984375 0.1875q0.921875 0.28125 1.578125 0.984375q0.671875 0.6875 1.015625 1.6875q0.34375 1.0 0.34375 2.484375q0 1.296875 -0.328125 2.234375q-0.390625 1.140625 -1.125 1.84375q-0.546875 0.546875 -1.484375 0.84375q-0.703125 0.21875 -1.875 0.21875l-3.984375 0l0 -10.484375zm2.109375 1.765625l0 6.953125l1.59375 0q0.875 0 1.28125 -0.109375q0.515625 -0.125 0.84375 -0.421875q0.34375 -0.3125 0.5625 -1.015625q0.21875 -0.703125 0.21875 -1.921875q0 -1.21875 -0.21875 -1.859375q-0.21875 -0.65625 -0.609375 -1.015625q-0.375 -0.375 -0.96875 -0.5q-0.453125 -0.109375 -1.75 -0.109375l-0.953125 0z" fill-rule="nonzero"/><path fill="#000000" fill-opacity="0.0" d="m237.65355 29.485565l69.76378 0l0 28.409447l-69.76378 0z" fill-rule="evenodd"/><path fill="#000000" d="m259.88397 45.704037l0 -2.25l5.8125 0l0 5.3125q-0.84375 0.828125 -2.453125 1.453125q-1.609375 0.625 -3.25 0.625q-2.09375 0 -3.65625 -0.875q-1.5625153 -0.890625 -2.3437653 -2.515625q-0.78125 -1.640625 -0.78125 -3.5625q0 -2.09375 0.875 -3.703125q0.875 -1.625 2.5625153 -2.5q1.28125 -0.65625 3.203125 -0.65625q2.484375 0 3.875 1.046875q1.40625 1.03125 1.796875 2.875l-2.671875 0.5q-0.28125 -0.984375 -1.0625 -1.546875q-0.78125 -0.578125 -1.9375 -0.578125q-1.78125 0 -2.828125 1.125q-1.0312653 1.125 -1.0312653 3.328125q0 2.375 1.0468903 3.5625q1.0625 1.1875 2.78125 1.1875q0.84375 0 1.6875 -0.328125q0.859375 -0.328125 1.46875 -0.8125l0 -1.6875l-3.09375 0zm8.339539 4.90625l0 -13.359375l2.703125 0l0 5.9375l5.4375 -5.9375l3.625 0l-5.015625 5.203125l5.296875 8.15625l-3.484375 0l-3.671875 -6.265625l-2.1875 2.234375l0 4.03125l-2.703125 0zm13.441711 0l0 -13.359375l9.90625 0l0 2.265625l-7.21875 0l0 2.953125l6.71875 0l0 2.25l-6.71875 0l0 3.640625l7.46875 0l0 2.25l-10.15625 0z" fill-rule="nonzero"/><path fill="#b7b7b7" d="m359.33032 83.24672l59.370087 0l0 16.598427l-59.370087 0z" fill-rule="evenodd"/><path stroke="#000000" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m359.33032 83.24672l59.370087 0l0 16.598427l-59.370087 0z" fill-rule="evenodd"/><path fill="#ffffff" d="m373.3694 93.922806l2.0 -0.203125q0.078125 0.6875 0.5 1.078125q0.421875 0.390625 0.984375 0.390625q0.625 0 1.0625 -0.5q0.4375 -0.515625 0.4375 -1.546875q0 -0.96875 -0.4375 -1.453125q-0.4375 -0.484375 -1.125 -0.484375q-0.859375 0 -1.546875 0.765625l-1.640625 -0.234375l1.03125 -5.453125l5.328125 0l0 1.875l-3.796875 0l-0.3125 1.78125q0.671875 -0.328125 1.375 -0.328125q1.328125 0 2.25 0.96875q0.9375 0.96875 0.9375 2.53125q0 1.296875 -0.75 2.3125q-1.015625 1.375 -2.84375 1.375q-1.453125 0 -2.375 -0.765625q-0.90625 -0.78125 -1.078125 -2.109375zm12.057373 2.703125l0 -2.109375l-4.296875 0l0 -1.765625l4.5625 -6.65625l1.6875 0l0 6.65625l1.296875 0l0 1.765625l-1.296875 0l0 2.109375l-1.953125 0zm0 -3.875l0 -3.59375l-2.40625 3.59375l2.40625 0zm4.1354675 1.09375l1.953125 -0.234375q0.09375 0.734375 0.5 1.140625q0.40625 0.390625 0.984375 0.390625q0.625 0 1.046875 -0.46875q0.421875 -0.484375 0.421875 -1.28125q0 -0.765625 -0.40625 -1.203125q-0.40625 -0.4375 -0.984375 -0.4375q-0.390625 0 -0.921875 0.140625l0.21875 -1.640625q0.8125 0.03125 1.234375 -0.34375q0.4375 -0.375 0.4375 -1.0q0 -0.53125 -0.3125 -0.84375q-0.3125 -0.3125 -0.84375 -0.3125q-0.515625 0 -0.875 0.359375q-0.359375 0.34375 -0.4375 1.03125l-1.859375 -0.3125q0.1875 -0.953125 0.578125 -1.515625q0.390625 -0.578125 1.078125 -0.890625q0.703125 -0.328125 1.578125 -0.328125q1.484375 0 2.375 0.9375q0.734375 0.78125 0.734375 1.75q0 1.375 -1.515625 2.203125q0.90625 0.1875 1.4375 0.875q0.546875 0.671875 0.546875 1.609375q0 1.390625 -1.015625 2.359375q-1.0 0.96875 -2.5 0.96875q-1.421875 0 -2.359375 -0.8125q-0.9375 -0.8125 -1.09375 -2.140625zm15.010498 0.90625l0 1.875l-7.046875 0q0.125 -1.0625 0.6875 -2.0q0.578125 -0.953125 2.265625 -2.515625q1.359375 -1.28125 1.671875 -1.71875q0.40625 -0.625 0.40625 -1.234375q0 -0.671875 -0.359375 -1.03125q-0.359375 -0.359375 -1.0 -0.359375q-0.625 0 -1.0 0.375q-0.375 0.375 -0.421875 1.25l-2.015625 -0.1875q0.1875 -1.671875 1.125 -2.390625q0.953125 -0.71875 2.375 -0.71875q1.546875 0 2.421875 0.84375q0.890625 0.828125 0.890625 2.078125q0 0.703125 -0.25 1.34375q-0.25 0.640625 -0.796875 1.34375q-0.375 0.46875 -1.328125 1.34375q-0.9375 0.859375 -1.203125 1.15625q-0.25 0.28125 -0.40625 0.546875l3.984375 0z" fill-rule="nonzero"/><path fill="#000000" fill-opacity="0.0" d="m373.18372 115.543304l91.87402 0l0 28.409454l-91.87402 0z" fill-rule="evenodd"/><path fill="#000000" d="m392.3106 117.00115l3.828125 0q1.125 0 1.671875 0.09375q0.5625 0.09375 1.0 0.40625q0.4375 0.296875 0.734375 0.796875q0.296875 0.5 0.296875 1.109375q0 0.671875 -0.359375 1.234375q-0.359375 0.5625 -0.984375 0.84375q0.875 0.25 1.34375 0.859375q0.46875 0.609375 0.46875 1.4375q0 0.65625 -0.3125 1.265625q-0.296875 0.609375 -0.828125 0.984375q-0.515625 0.359375 -1.28125 0.453125q-0.484375 0.046875 -2.328125 0.0625l-3.25 0l0 -9.546875zm1.9375 1.59375l0 2.203125l1.265625 0q1.125 0 1.390625 -0.03125q0.5 -0.0625 0.78125 -0.34375q0.28125 -0.28125 0.28125 -0.734375q0 -0.453125 -0.25 -0.71875q-0.234375 -0.28125 -0.71875 -0.34375q-0.28125 -0.03125 -1.640625 -0.03125l-1.109375 0zm0 3.796875l0 2.546875l1.78125 0q1.046875 0 1.3125 -0.0625q0.4375 -0.078125 0.703125 -0.375q0.265625 -0.296875 0.265625 -0.8125q0 -0.421875 -0.203125 -0.71875q-0.203125 -0.3125 -0.609375 -0.4375q-0.390625 -0.140625 -1.703125 -0.140625l-1.546875 0zm9.047028 -0.640625l-1.671875 -0.3125q0.28125 -1.0 0.96875 -1.46875q0.6875 -0.484375 2.03125 -0.484375q1.21875 0 1.8125 0.296875q0.609375 0.28125 0.84375 0.734375q0.25 0.4375 0.25 1.625l-0.015625 2.140625q0 0.90625 0.078125 1.34375q0.09375 0.421875 0.34375 0.921875l-1.8125 0q-0.078125 -0.1875 -0.171875 -0.546875q-0.046875 -0.15625 -0.078125 -0.203125q-0.46875 0.453125 -1.0 0.6875q-0.53125 0.21875 -1.140625 0.21875q-1.0625 0 -1.6875 -0.578125q-0.609375 -0.578125 -0.609375 -1.46875q0 -0.578125 0.28125 -1.03125q0.28125 -0.46875 0.78125 -0.703125q0.515625 -0.25 1.46875 -0.4375q1.28125 -0.234375 1.765625 -0.453125l0 -0.171875q0 -0.53125 -0.265625 -0.75q-0.25 -0.234375 -0.96875 -0.234375q-0.484375 0 -0.765625 0.203125q-0.265625 0.1875 -0.4375 0.671875zm2.4375 1.484375q-0.34375 0.109375 -1.109375 0.28125q-0.765625 0.15625 -1.0 0.3125q-0.359375 0.25 -0.359375 0.640625q0 0.390625 0.28125 0.671875q0.296875 0.28125 0.734375 0.28125q0.5 0 0.953125 -0.328125q0.328125 -0.25 0.4375 -0.609375q0.0625 -0.234375 0.0625 -0.890625l0 -0.359375zm9.631226 -1.5625l-1.796875 0.328125q-0.09375 -0.53125 -0.421875 -0.796875q-0.328125 -0.28125 -0.828125 -0.28125q-0.6875 0 -1.09375 0.46875q-0.40625 0.46875 -0.40625 1.578125q0 1.234375 0.40625 1.75q0.421875 0.5 1.109375 0.5q0.53125 0 0.859375 -0.296875q0.328125 -0.296875 0.46875 -1.03125l1.796875 0.3125q-0.28125 1.234375 -1.078125 1.875q-0.796875 0.625 -2.125 0.625q-1.515625 0 -2.421875 -0.953125q-0.90625 -0.953125 -0.90625 -2.65625q0 -1.703125 0.90625 -2.65625q0.90625 -0.953125 2.453125 -0.953125q1.265625 0 2.0 0.546875q0.75 0.53125 1.078125 1.640625zm1.3187256 4.875l0 -9.546875l1.828125 0l0 5.0625l2.140625 -2.421875l2.25 0l-2.359375 2.515625l2.53125 4.390625l-1.96875 0l-1.734375 -3.109375l-0.859375 0.890625l0 2.21875l-1.828125 0zm11.474976 -2.203125l1.828125 0.3125q-0.34375 1.0 -1.109375 1.53125q-0.765625 0.515625 -1.890625 0.515625q-1.8125 0 -2.671875 -1.171875q-0.6875 -0.953125 -0.6875 -2.390625q0 -1.71875 0.890625 -2.6875q0.90625 -0.96875 2.28125 -0.96875q1.546875 0 2.4375 1.015625q0.890625 1.015625 0.84375 3.125l-4.578125 0q0.015625 0.8125 0.4375 1.265625q0.421875 0.453125 1.0625 0.453125q0.421875 0 0.71875 -0.234375q0.296875 -0.234375 0.4375 -0.765625zm0.109375 -1.84375q-0.015625 -0.796875 -0.40625 -1.203125q-0.390625 -0.421875 -0.953125 -0.421875q-0.59375 0 -0.984375 0.4375q-0.390625 0.4375 -0.390625 1.1875l2.734375 0zm9.599976 4.046875l-1.828125 0l0 -3.53125q0 -1.109375 -0.125 -1.4375q-0.109375 -0.328125 -0.375 -0.515625q-0.265625 -0.1875 -0.640625 -0.1875q-0.46875 0 -0.859375 0.265625q-0.375 0.265625 -0.515625 0.703125q-0.140625 0.421875 -0.140625 1.578125l0 3.125l-1.828125 0l0 -6.90625l1.703125 0l0 1.015625q0.90625 -1.171875 2.28125 -1.171875q0.609375 0 1.109375 0.21875q0.5 0.21875 0.75 0.5625q0.265625 0.328125 0.359375 0.765625q0.109375 0.421875 0.109375 1.21875l0 4.296875zm8.188232 0l-1.703125 0l0 -1.015625q-0.421875 0.59375 -1.0 0.890625q-0.578125 0.28125 -1.15625 0.28125q-1.1875 0 -2.046875 -0.953125q-0.84375 -0.96875 -0.84375 -2.6875q0 -1.75 0.828125 -2.65625q0.828125 -0.921875 2.09375 -0.921875q1.15625 0 2.0 0.953125l0 -3.4375l1.828125 0l0 9.546875zm-4.875 -3.609375q0 1.109375 0.296875 1.609375q0.4375 0.71875 1.234375 0.71875q0.640625 0 1.078125 -0.53125q0.4375 -0.546875 0.4375 -1.609375q0 -1.203125 -0.4375 -1.71875q-0.421875 -0.53125 -1.09375 -0.53125q-0.640625 0 -1.078125 0.515625q-0.4375 0.515625 -0.4375 1.546875z" fill-rule="nonzero"/><path fill="#000000" d="m395.89038 139.43866l1.875 -0.171875q0.171875 0.9375 0.6875 1.390625q0.515625 0.4375 1.390625 0.4375q0.9375 0 1.40625 -0.390625q0.46875 -0.40625 0.46875 -0.921875q0 -0.34375 -0.203125 -0.578125q-0.1875 -0.234375 -0.6875 -0.421875q-0.34375 -0.109375 -1.546875 -0.40625q-1.546875 -0.390625 -2.171875 -0.953125q-0.875 -0.78125 -0.875 -1.921875q0 -0.71875 0.40625 -1.359375q0.421875 -0.640625 1.1875 -0.96875q0.78125 -0.328125 1.890625 -0.328125q1.796875 0 2.703125 0.796875q0.90625 0.78125 0.953125 2.09375l-1.921875 0.078125q-0.125 -0.734375 -0.53125 -1.046875q-0.40625 -0.328125 -1.234375 -0.328125q-0.828125 0 -1.3125 0.34375q-0.296875 0.21875 -0.296875 0.59375q0 0.34375 0.28125 0.578125q0.359375 0.3125 1.765625 0.640625q1.40625 0.328125 2.078125 0.6875q0.6875 0.359375 1.0625 0.96875q0.375 0.609375 0.375 1.515625q0 0.828125 -0.453125 1.546875q-0.453125 0.703125 -1.296875 1.0625q-0.828125 0.34375 -2.0625 0.34375q-1.8125 0 -2.78125 -0.828125q-0.96875 -0.84375 -1.15625 -2.453125zm13.3585205 0.90625l1.828125 0.3125q-0.34375 1.0 -1.109375 1.53125q-0.765625 0.515625 -1.890625 0.515625q-1.8125 0 -2.671875 -1.171875q-0.6875 -0.953125 -0.6875 -2.390625q0 -1.71875 0.890625 -2.6875q0.90625 -0.96875 2.28125 -0.96875q1.546875 0 2.4375 1.015625q0.890625 1.015625 0.84375 3.125l-4.578125 0q0.015625 0.8125 0.4375 1.265625q0.421875 0.453125 1.0625 0.453125q0.421875 0 0.71875 -0.234375q0.296875 -0.234375 0.4375 -0.765625zm0.109375 -1.84375q-0.015625 -0.796875 -0.40625 -1.203125q-0.390625 -0.421875 -0.953125 -0.421875q-0.59375 0 -0.984375 0.4375q-0.390625 0.4375 -0.390625 1.1875l2.734375 0zm5.0531006 4.046875l-1.828125 0l0 -6.90625l1.703125 0l0 0.984375q0.4375 -0.703125 0.78125 -0.921875q0.34375 -0.21875 0.796875 -0.21875q0.625 0 1.203125 0.34375l-0.5625 1.59375q-0.46875 -0.296875 -0.859375 -0.296875q-0.390625 0 -0.65625 0.21875q-0.265625 0.203125 -0.421875 0.765625q-0.15625 0.546875 -0.15625 2.296875l0 2.140625zm5.3430176 0l-2.78125 -6.90625l1.90625 0l1.3125 3.515625l0.375 1.1875q0.140625 -0.453125 0.1875 -0.59375q0.09375 -0.296875 0.1875 -0.59375l1.328125 -3.515625l1.875 0l-2.75 6.90625l-1.640625 0zm5.5062256 -7.84375l0 -1.703125l1.828125 0l0 1.703125l-1.828125 0zm0 7.84375l0 -6.90625l1.828125 0l0 6.90625l-1.828125 0zm9.734253 -4.875l-1.796875 0.328125q-0.09375 -0.53125 -0.421875 -0.796875q-0.328125 -0.28125 -0.828125 -0.28125q-0.6875 0 -1.09375 0.46875q-0.40625 0.46875 -0.40625 1.578125q0 1.234375 0.40625 1.75q0.421875 0.5 1.109375 0.5q0.53125 0 0.859375 -0.296875q0.328125 -0.296875 0.46875 -1.03125l1.796875 0.3125q-0.28125 1.234375 -1.078125 1.875q-0.796875 0.625 -2.125 0.625q-1.515625 0 -2.421875 -0.953125q-0.90625 -0.953125 -0.90625 -2.65625q0 -1.703125 0.90625 -2.65625q0.90625 -0.953125 2.453125 -0.953125q1.265625 0 2.0 0.546875q0.75 0.53125 1.078125 1.640625zm5.3812256 2.671875l1.828125 0.3125q-0.34375 1.0 -1.109375 1.53125q-0.765625 0.515625 -1.890625 0.515625q-1.8125 0 -2.671875 -1.171875q-0.6875 -0.953125 -0.6875 -2.390625q0 -1.71875 0.890625 -2.6875q0.90625 -0.96875 2.28125 -0.96875q1.546875 0 2.4375 1.015625q0.890625 1.015625 0.84375 3.125l-4.578125 0q0.015625 0.8125 0.4375 1.265625q0.421875 0.453125 1.0625 0.453125q0.421875 0 0.71875 -0.234375q0.296875 -0.234375 0.4375 -0.765625zm0.109375 -1.84375q-0.015625 -0.796875 -0.40625 -1.203125q-0.390625 -0.421875 -0.953125 -0.421875q-0.59375 0 -0.984375 0.4375q-0.390625 0.4375 -0.390625 1.1875l2.734375 0z" fill-rule="nonzero"/><path fill="#b7b7b7" d="m280.83466 115.105156l0 164.91339l-16.59845 0l0 -164.91339z" fill-rule="evenodd"/><path stroke="#000000" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m280.83466 115.105156l0 164.91339l-16.59845 0l0 -164.91339z" fill-rule="evenodd"/><path fill="#ffffff" d="m267.45544 139.63248l10.40625 0l0 2.109375l-8.640625 0l0 5.28125l-1.765625 0l0 -7.390625zm3.90625 8.421371q1.0 0 1.9375 0.484375q0.9375 0.5 1.421875 1.40625q0.5 0.90625 0.5 2.015625q0 1.71875 -1.125 2.828125q-1.109375 1.109375 -2.8125 1.109375q-1.734375 0 -2.875 -1.125q-1.125 -1.109375 -1.125 -2.796875q0 -1.046875 0.46875 -2.0q0.46875 -0.9375 1.375 -1.4375q0.921875 -0.484375 2.234375 -0.484375zm-0.109375 2.046875q-1.125 0 -1.734375 0.53125q-0.59375 0.546875 -0.59375 1.34375q0 0.78125 0.59375 1.3125q0.609375 0.53125 1.75 0.53125q1.125 0 1.71875 -0.53125q0.609375 -0.53125 0.609375 -1.3125q0 -0.796875 -0.609375 -1.34375q-0.59375 -0.53125 -1.734375 -0.53125zm1.484375 8.874496l0.328125 -1.828125q1.109375 0.296875 1.625 1.046875q0.53125 0.765625 0.53125 2.234375q0 1.359375 -0.3125 2.015625q-0.3125 0.65625 -0.8125 0.921875q-0.484375 0.265625 -1.796875 0.265625l-2.34375 -0.015625q-1.0 0 -1.484375 0.09375q-0.46875 0.09375 -1.015625 0.359375l0 -1.984375q0.203125 -0.078125 0.59375 -0.203125q0.171875 -0.046875 0.234375 -0.0625q-0.5 -0.515625 -0.75 -1.109375q-0.25 -0.578125 -0.25 -1.25q0 -1.171875 0.640625 -1.859375q0.640625 -0.671875 1.609375 -0.671875q0.640625 0 1.140625 0.3125q0.515625 0.3125 0.78125 0.859375q0.265625 0.5625 0.46875 1.609375q0.265625 1.40625 0.484375 1.953125l0.203125 0q0.578125 0 0.828125 -0.296875q0.25 -0.28125 0.25 -1.078125q0 -0.53125 -0.21875 -0.84375q-0.203125 -0.296875 -0.734375 -0.46875zm-1.625 2.6875q-0.140625 -0.390625 -0.328125 -1.234375q-0.171875 -0.828125 -0.34375 -1.09375q-0.265625 -0.390625 -0.703125 -0.390625q-0.421875 0 -0.734375 0.3125q-0.296875 0.3125 -0.296875 0.8125q0 0.53125 0.359375 1.03125q0.265625 0.359375 0.65625 0.484375q0.265625 0.078125 0.984375 0.078125l0.40625 0zm-3.65625 10.916733l0 -1.859375l1.109375 0q-0.640625 -0.46875 -0.96875 -1.109375q-0.3125 -0.625 -0.3125 -1.265625q0 -1.3125 1.046875 -2.25q1.0625 -0.9375 2.953125 -0.9375q1.9375 0 2.9375 0.90625q1.0 0.921875 1.0 2.3125q0 1.265625 -1.046875 2.203125l3.765625 0l0 2.0l-10.484375 0zm3.96875 -5.359375q-1.21875 0 -1.765625 0.328125q-0.78125 0.5 -0.78125 1.359375q0 0.703125 0.578125 1.1875q0.59375 0.484375 1.765625 0.484375q1.3125 0 1.890625 -0.46875q0.578125 -0.46875 0.578125 -1.203125q0 -0.71875 -0.578125 -1.203125q-0.5625 -0.484375 -1.6875 -0.484375zm6.515625 7.3744965l0 4.1875q0 1.25 -0.109375 1.859375q-0.09375 0.609375 -0.421875 1.09375q-0.328125 0.484375 -0.875 0.796875q-0.546875 0.328125 -1.234375 0.328125q-0.734375 0 -1.359375 -0.40625q-0.609375 -0.390625 -0.90625 -1.0625q-0.28125 0.953125 -0.953125 1.46875q-0.671875 0.515625 -1.578125 0.515625q-0.71875 0 -1.40625 -0.328125q-0.671875 -0.328125 -1.078125 -0.90625q-0.390625 -0.578125 -0.484375 -1.421875q-0.0625 -0.53125 -0.078125 -2.5625l0 -3.5625l10.484375 0zm-1.734375 2.109375l-2.4375 0l0 1.390625q0 1.234375 0.046875 1.546875q0.0625 0.53125 0.375 0.84375q0.3125 0.3125 0.8125 0.3125q0.484375 0 0.78125 -0.265625q0.3125 -0.265625 0.375 -0.796875q0.046875 -0.3125 0.046875 -1.8125l0 -1.21875zm-4.171875 0l-2.8125 0l0 1.96875q0 1.140625 0.0625 1.453125q0.09375 0.46875 0.421875 0.765625q0.328125 0.296875 0.890625 0.296875q0.46875 0 0.796875 -0.234375q0.328125 -0.21875 0.484375 -0.65625q0.15625 -0.4375 0.15625 -1.875l0 -1.71875zm0.703125 9.959274l0.328125 -1.828125q1.109375 0.296875 1.625 1.046875q0.53125 0.765625 0.53125 2.234375q0 1.359375 -0.3125 2.015625q-0.3125 0.65625 -0.8125 0.921875q-0.484375 0.265625 -1.796875 0.265625l-2.34375 -0.015625q-1.0 0 -1.484375 0.09375q-0.46875 0.09375 -1.015625 0.359375l0 -1.984375q0.203125 -0.078125 0.59375 -0.203125q0.171875 -0.046875 0.234375 -0.0625q-0.5 -0.515625 -0.75 -1.109375q-0.25 -0.578125 -0.25 -1.25q0 -1.171875 0.640625 -1.859375q0.640625 -0.671875 1.609375 -0.671875q0.640625 0 1.140625 0.3125q0.515625 0.3125 0.78125 0.859375q0.265625 0.5625 0.46875 1.609375q0.265625 1.40625 0.484375 1.953125l0.203125 0q0.578125 0 0.828125 -0.296875q0.25 -0.28125 0.25 -1.078125q0 -0.53125 -0.21875 -0.84375q-0.203125 -0.296875 -0.734375 -0.46875zm-1.625 2.6875q-0.140625 -0.390625 -0.328125 -1.234375q-0.171875 -0.828125 -0.34375 -1.09375q-0.265625 -0.390625 -0.703125 -0.390625q-0.421875 0 -0.734375 0.3125q-0.296875 0.3125 -0.296875 0.8125q0 0.53125 0.359375 1.03125q0.265625 0.359375 0.65625 0.484375q0.265625 0.078125 0.984375 0.078125l0.40625 0zm-3.65625 3.9479828l10.484375 0l0 2.015625l-10.484375 0l0 -2.015625zm5.28125 5.5876007l0.328125 -1.828125q1.109375 0.296875 1.625 1.046875q0.53125 0.765625 0.53125 2.234375q0 1.359375 -0.3125 2.015625q-0.3125 0.65625 -0.8125 0.921875q-0.484375 0.265625 -1.796875 0.265625l-2.34375 -0.015625q-1.0 0 -1.484375 0.09375q-0.46875 0.09375 -1.015625 0.359375l0 -1.984375q0.203125 -0.078125 0.59375 -0.203125q0.171875 -0.046875 0.234375 -0.0625q-0.5 -0.515625 -0.75 -1.109375q-0.25 -0.578125 -0.25 -1.25q0 -1.171875 0.640625 -1.859375q0.640625 -0.671875 1.609375 -0.671875q0.640625 0 1.140625 0.3125q0.515625 0.3125 0.78125 0.859375q0.265625 0.5625 0.46875 1.609375q0.265625 1.40625 0.484375 1.953125l0.203125 0q0.578125 0 0.828125 -0.296875q0.25 -0.28125 0.25 -1.078125q0 -0.53125 -0.21875 -0.84375q-0.203125 -0.296875 -0.734375 -0.46875zm-1.625 2.6875q-0.140625 -0.390625 -0.328125 -1.234375q-0.171875 -0.828125 -0.34375 -1.09375q-0.265625 -0.390625 -0.703125 -0.390625q-0.421875 0 -0.734375 0.3125q-0.296875 0.3125 -0.296875 0.8125q0 0.53125 0.359375 1.03125q0.265625 0.359375 0.65625 0.484375q0.265625 0.078125 0.984375 0.078125l0.40625 0zm-3.65625 10.869858l0 -2.015625l3.875 0q1.234375 0 1.59375 -0.125q0.359375 -0.125 0.5625 -0.421875q0.203125 -0.296875 0.203125 -0.703125q0 -0.515625 -0.296875 -0.9375q-0.28125 -0.40625 -0.75 -0.5625q-0.46875 -0.15625 -1.75 -0.15625l-3.4375 0l0 -2.015625l7.59375 0l0 1.875l-1.109375 0q1.28125 1.0 1.28125 2.5q0 0.671875 -0.234375 1.21875q-0.234375 0.546875 -0.609375 0.828125q-0.375 0.296875 -0.84375 0.40625q-0.46875 0.109375 -1.359375 0.109375l-4.71875 0zm5.359375 8.655746l-0.359375 -1.96875q0.59375 -0.109375 0.890625 -0.46875q0.296875 -0.34375 0.296875 -0.90625q0 -0.75 -0.515625 -1.203125q-0.515625 -0.453125 -1.734375 -0.453125q-1.359375 0 -1.921875 0.453125q-0.546875 0.453125 -0.546875 1.21875q0 0.578125 0.328125 0.9375q0.328125 0.375 1.109375 0.515625l-0.328125 1.984375q-1.359375 -0.3125 -2.0625 -1.1875q-0.6875 -0.875 -0.6875 -2.328125q0 -1.671875 1.046875 -2.671875q1.0625 -0.984375 2.921875 -0.984375q1.875 0 2.921875 0.984375q1.046875 1.0 1.046875 2.703125q0 1.390625 -0.59375 2.203125q-0.59375 0.828125 -1.8125 1.171875zm-2.9375 5.932358l-0.34375 2.0q-1.09375 -0.375 -1.671875 -1.21875q-0.578125 -0.828125 -0.578125 -2.078125q0 -1.984375 1.296875 -2.9375q1.03125 -0.75 2.625 -0.75q1.875 0 2.9375 0.984375q1.078125 0.984375 1.078125 2.5q0 1.6875 -1.125 2.671875q-1.109375 0.984375 -3.421875 0.9375l0 -5.03125q-0.890625 0.015625 -1.390625 0.484375q-0.5 0.46875 -0.5 1.15625q0 0.46875 0.25 0.78125q0.265625 0.328125 0.84375 0.5zm2.03125 0.109375q0.875 -0.015625 1.328125 -0.453125q0.453125 -0.421875 0.453125 -1.03125q0 -0.671875 -0.484375 -1.09375q-0.46875 -0.421875 -1.296875 -0.421875l0 3.0zm-4.453125 5.572983l0 -2.015625l7.59375 0l0 1.859375l-1.078125 0q0.765625 0.484375 1.0 0.859375q0.25 0.390625 0.25 0.875q0 0.6875 -0.375 1.328125l-1.75 -0.625q0.328125 -0.5 0.328125 -0.9375q0 -0.421875 -0.234375 -0.71875q-0.234375 -0.296875 -0.84375 -0.46875q-0.609375 -0.15625 -2.546875 -0.15625l-2.34375 0zm5.59375 4.1567535l2.0 0l0 2.015625l-2.0 0l0 -2.015625zm-5.59375 0l2.015625 0l0 2.015625l-2.015625 0l0 -2.015625zm5.65625 5.7868958q0.328125 -0.78125 0.90625 -1.140625q0.578125 -0.34375 1.265625 -0.34375q1.171875 0 1.9375 0.8125q0.765625 0.828125 0.765625 2.328125q0 1.5 -0.765625 2.328125q-0.765625 0.828125 -1.9375 0.828125q-0.734375 0 -1.296875 -0.390625q-0.5625 -0.375 -0.875 -1.0625q-0.34375 0.875 -1.015625 1.328125q-0.671875 0.46875 -1.5625 0.46875q-1.453125 0 -2.359375 -0.9375q-0.90625 -0.921875 -0.90625 -2.46875q0 -1.421875 0.75 -2.375q0.890625 -1.125 2.4375 -1.125q0.84375 0 1.5625 0.421875q0.71875 0.421875 1.09375 1.328125zm2.03125 0.421875q-0.609375 0 -0.953125 0.328125q-0.328125 0.34375 -0.328125 0.90625q0 0.578125 0.34375 0.921875q0.34375 0.34375 0.9375 0.34375q0.578125 0 0.90625 -0.34375q0.34375 -0.328125 0.34375 -0.890625q0 -0.578125 -0.34375 -0.921875q-0.34375 -0.34375 -0.90625 -0.34375zm-4.5 -0.1875q-0.828125 0 -1.296875 0.421875q-0.453125 0.421875 -0.453125 1.0625q0 0.625 0.4375 1.03125q0.453125 0.40625 1.296875 0.40625q0.734375 0 1.171875 -0.421875q0.453125 -0.40625 0.453125 -1.046875q0 -0.734375 -0.515625 -1.09375q-0.5 -0.359375 -1.09375 -0.359375zm7.34375 9.588608q0 1.53125 -1.078125 2.390625q-1.296875 1.015625 -4.28125 1.015625q-2.96875 0 -4.28125 -1.03125q-1.0625 -0.84375 -1.0625 -2.375q0 -1.53125 1.171875 -2.46875q1.171875 -0.9375 4.203125 -0.9375q2.953125 0 4.25 1.03125q1.078125 0.859375 1.078125 2.375zm-1.671875 0q0 -0.359375 -0.234375 -0.65625q-0.21875 -0.28125 -0.828125 -0.4375q-0.78125 -0.203125 -2.625 -0.203125q-1.84375 0 -2.53125 0.1875q-0.6875 0.1875 -0.921875 0.46875q-0.234375 0.28125 -0.234375 0.640625q0 0.375 0.234375 0.65625q0.234375 0.28125 0.84375 0.4375q0.765625 0.21875 2.609375 0.21875q1.84375 0 2.53125 -0.1875q0.703125 -0.1875 0.921875 -0.46875q0.234375 -0.28125 0.234375 -0.65625z" fill-rule="nonzero"/><path fill="#b7b7b7" d="m315.073 189.18855l38.614166 0l0 16.598434l-38.614166 0z" fill-rule="evenodd"/><path stroke="#000000" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m315.073 189.18855l38.614166 0l0 16.598434l-38.614166 0z" fill-rule="evenodd"/><path fill="#ffffff" d="m328.57272 196.91151q-0.78125 -0.328125 -1.140625 -0.90625q-0.34375 -0.578125 -0.34375 -1.265625q0 -1.171875 0.8125 -1.9375q0.828125 -0.765625 2.328125 -0.765625q1.5 0 2.328125 0.765625q0.828125 0.765625 0.828125 1.9375q0 0.734375 -0.390625 1.296875q-0.375 0.5625 -1.0625 0.875q0.875 0.34375 1.328125 1.015625q0.46875 0.671875 0.46875 1.5625q0 1.453125 -0.9375 2.359375q-0.921875 0.90625 -2.46875 0.90625q-1.421875 0 -2.375 -0.75q-1.125 -0.890625 -1.125 -2.4375q0 -0.84375 0.421875 -1.5625q0.421875 -0.71875 1.328125 -1.09375zm0.421875 -2.03125q0 0.609375 0.328125 0.953125q0.34375 0.328125 0.90625 0.328125q0.578125 0 0.921875 -0.34375q0.34375 -0.34375 0.34375 -0.9375q0 -0.578125 -0.34375 -0.90625q-0.328125 -0.34375 -0.890625 -0.34375q-0.578125 0 -0.921875 0.34375q-0.34375 0.34375 -0.34375 0.90625zm-0.1875 4.5q0 0.828125 0.421875 1.296875q0.421875 0.453125 1.0625 0.453125q0.625 0 1.03125 -0.4375q0.40625 -0.453125 0.40625 -1.296875q0 -0.734375 -0.421875 -1.171875q-0.40625 -0.453125 -1.046875 -0.453125q-0.734375 0 -1.09375 0.515625q-0.359375 0.5 -0.359375 1.09375zm9.588593 -7.34375q1.53125 0 2.390625 1.078125q1.015625 1.296875 1.015625 4.28125q0 2.96875 -1.03125 4.28125q-0.84375 1.0625 -2.375 1.0625q-1.53125 0 -2.46875 -1.171875q-0.9375 -1.171875 -0.9375 -4.203125q0 -2.953125 1.03125 -4.25q0.859375 -1.078125 2.375 -1.078125zm0 1.671875q-0.359375 0 -0.65625 0.234375q-0.28125 0.21875 -0.4375 0.828125q-0.203125 0.78125 -0.203125 2.625q0 1.84375 0.1875 2.53125q0.1875 0.6875 0.46875 0.921875q0.28125 0.234375 0.640625 0.234375q0.375 0 0.65625 -0.234375q0.28125 -0.234375 0.4375 -0.84375q0.21875 -0.765625 0.21875 -2.609375q0 -1.84375 -0.1875 -2.53125q-0.1875 -0.703125 -0.46875 -0.921875q-0.28125 -0.234375 -0.65625 -0.234375z" fill-rule="nonzero"/><path fill="#000000" fill-opacity="0.0" d="m389.01535 99.845146l0.18899536 59.811028" fill-rule="evenodd"/><path stroke="#000000" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m389.0273 103.63513l0.16503906 52.23105" fill-rule="evenodd"/><path fill="#000000" stroke="#000000" stroke-width="1.0" stroke-linecap="butt" d="m389.0169 100.34514c0.9085083 -0.0028686523 1.6473389 0.7312927 1.6502075 1.639801c0.0028686523 0.9085007 -0.7312927 1.647316 -1.639801 1.6501846c-0.9085083 0.0028686523 -1.6473083 -0.7312927 -1.650177 -1.6397934c-0.0028686523 -0.9085007 0.7312927 -1.647316 1.6397705 -1.6501923z" fill-rule="nonzero"/><path fill="#000000" stroke="#000000" stroke-width="1.0" stroke-linecap="butt" d="m389.20273 159.15617c-0.9085083 0.0028686523 -1.6473083 -0.7312927 -1.650177 -1.639801c-0.0028686523 -0.90849304 0.7312927 -1.6473083 1.639801 -1.6501923c0.9084778 -0.0028686523 1.6473083 0.7312927 1.650177 1.639801c0.0028686523 0.9085083 -0.7312927 1.6473236 -1.639801 1.6501923z" fill-rule="nonzero"/><path fill="#000000" fill-opacity="0.0" d="m382.15955 298.32547l47.994965 0l0 42.897736l-47.994965 0z" fill-rule="evenodd"/><g transform="matrix(0.024997375328083986 0.0 0.0 0.024998687664041994 382.15955853018374 298.32546771653546)"><clipPath id="p.2"><path d="m0 0l1920.0 0l0 1716.0l-1920.0 0z" clip-rule="evenodd"/></clipPath><image clip-path="url(#p.2)" fill="#000" width="1920.0" height="1716.0" x="0.0" y="0.0" preserveAspectRatio="none" xlink:href="data:image/png;base64,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"/></g><path fill="#000000" fill-opacity="0.0" d="m389.20734 194.13838l-0.56692505 80.66142" fill-rule="evenodd"/><path stroke="#000000" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m389.1807 197.92827l-0.51361084 73.08162" fill-rule="evenodd"/><path fill="#000000" stroke="#000000" stroke-width="1.0" stroke-linecap="butt" d="m389.20383 194.63835c0.9085083 0.0063934326 1.639801 0.7480469 1.6333923 1.6565247c-0.006378174 0.90849304 -0.74801636 1.6397858 -1.6564941 1.6333923c-0.9085083 -0.006378174 -1.639801 -0.7480316 -1.6334229 -1.6565094c0.0064086914 -0.90849304 0.7480469 -1.6397858 1.6565247 -1.6334076z" fill-rule="nonzero"/><path fill="#000000" stroke="#000000" stroke-width="1.0" stroke-linecap="butt" d="m388.64392 274.2998c-0.9084778 -0.006378174 -1.6397705 -0.7480469 -1.6333923 -1.6565247c0.006378174 -0.9084778 0.7480469 -1.6397705 1.6565247 -1.6333923c0.9084778 0.006378174 1.639801 0.74801636 1.6333923 1.6565247c-0.006378174 0.9084778 -0.74801636 1.6397705 -1.6565247 1.6333923z" fill-rule="nonzero"/><path fill="#000000" fill-opacity="0.0" d="m382.1601 223.26772l73.921265 0l0 16.59842l-73.921265 0z" fill-rule="evenodd"/><path fill="#000000" d="m405.37137 228.36693l-2.09375 0l-0.84375 -2.171875l-3.8125 0l-0.78125 2.171875l-2.046875 0l3.71875 -9.546875l2.03125 0l3.828125 9.546875zm-3.546875 -3.78125l-1.3125 -3.53125l-1.296875 3.53125l2.609375 0zm6.2970276 3.78125l-1.828125 0l0 -6.90625l1.703125 0l0 0.984375q0.4375 -0.703125 0.78125 -0.921875q0.34375 -0.21875 0.796875 -0.21875q0.625 0 1.203125 0.34375l-0.5625 1.59375q-0.46875 -0.296875 -0.859375 -0.296875q-0.390625 0 -0.65625 0.21875q-0.265625 0.203125 -0.421875 0.765625q-0.15625 0.546875 -0.15625 2.296875l0 2.140625zm6.608673 -6.90625l0 1.453125l-1.25 0l0 2.78125q0 0.84375 0.03125 0.984375q0.046875 0.140625 0.171875 0.234375q0.125 0.09375 0.3125 0.09375q0.25 0 0.734375 -0.171875l0.15625 1.421875q-0.640625 0.265625 -1.453125 0.265625q-0.484375 0 -0.890625 -0.15625q-0.390625 -0.171875 -0.578125 -0.4375q-0.1875 -0.265625 -0.265625 -0.71875q-0.046875 -0.3125 -0.046875 -1.28125l0 -3.015625l-0.84375 0l0 -1.453125l0.84375 0l0 -1.375l1.828125 -1.078125l0 2.453125l1.25 0zm1.2664795 -0.9375l0 -1.703125l1.828125 0l0 1.703125l-1.828125 0zm0 7.84375l0 -6.90625l1.828125 0l0 6.90625l-1.828125 0zm2.906128 -6.90625l1.015625 0l0 -0.53125q0 -0.859375 0.1875 -1.28125q0.1875 -0.4375 0.671875 -0.703125q0.5 -0.28125 1.265625 -0.28125q0.78125 0 1.53125 0.234375l-0.25 1.28125q-0.4375 -0.109375 -0.84375 -0.109375q-0.390625 0 -0.5625 0.1875q-0.171875 0.1875 -0.171875 0.703125l0 0.5l1.359375 0l0 1.4375l-1.359375 0l0 5.46875l-1.828125 0l0 -5.46875l-1.015625 0l0 -1.4375zm6.6102295 2.109375l-1.671875 -0.3125q0.28125 -1.0 0.96875 -1.46875q0.6875 -0.484375 2.03125 -0.484375q1.21875 0 1.8125 0.296875q0.609375 0.28125 0.84375 0.734375q0.25 0.4375 0.25 1.625l-0.015625 2.140625q0 0.90625 0.078125 1.34375q0.09375 0.421875 0.34375 0.921875l-1.8125 0q-0.078125 -0.1875 -0.171875 -0.546875q-0.046875 -0.15625 -0.078125 -0.203125q-0.46875 0.453125 -1.0 0.6875q-0.53125 0.21875 -1.140625 0.21875q-1.0625 0 -1.6875 -0.578125q-0.609375 -0.578125 -0.609375 -1.46875q0 -0.578125 0.28125 -1.03125q0.28125 -0.46875 0.78125 -0.703125q0.515625 -0.25 1.46875 -0.4375q1.28125 -0.234375 1.765625 -0.453125l0 -0.171875q0 -0.53125 -0.265625 -0.75q-0.25 -0.234375 -0.96875 -0.234375q-0.484375 0 -0.765625 0.203125q-0.265625 0.1875 -0.4375 0.671875zm2.4375 1.484375q-0.34375 0.109375 -1.109375 0.28125q-0.765625 0.15625 -1.0 0.3125q-0.359375 0.25 -0.359375 0.640625q0 0.390625 0.28125 0.671875q0.296875 0.28125 0.734375 0.28125q0.5 0 0.953125 -0.328125q0.328125 -0.25 0.4375 -0.609375q0.0625 -0.234375 0.0625 -0.890625l0 -0.359375zm9.631226 -1.5625l-1.796875 0.328125q-0.09375 -0.53125 -0.421875 -0.796875q-0.328125 -0.28125 -0.828125 -0.28125q-0.6875 0 -1.09375 0.46875q-0.40625 0.46875 -0.40625 1.578125q0 1.234375 0.40625 1.75q0.421875 0.5 1.109375 0.5q0.53125 0 0.859375 -0.296875q0.328125 -0.296875 0.46875 -1.03125l1.796875 0.3125q-0.28125 1.234375 -1.078125 1.875q-0.796875 0.625 -2.125 0.625q-1.515625 0 -2.421875 -0.953125q-0.90625 -0.953125 -0.90625 -2.65625q0 -1.703125 0.90625 -2.65625q0.90625 -0.953125 2.453125 -0.953125q1.265625 0 2.0 0.546875q0.75 0.53125 1.078125 1.640625zm4.5531006 -2.03125l0 1.453125l-1.25 0l0 2.78125q0 0.84375 0.03125 0.984375q0.046875 0.140625 0.171875 0.234375q0.125 0.09375 0.3125 0.09375q0.25 0 0.734375 -0.171875l0.15625 1.421875q-0.640625 0.265625 -1.453125 0.265625q-0.484375 0 -0.890625 -0.15625q-0.390625 -0.171875 -0.578125 -0.4375q-0.1875 -0.265625 -0.265625 -0.71875q-0.046875 -0.3125 -0.046875 -1.28125l0 -3.015625l-0.84375 0l0 -1.453125l0.84375 0l0 -1.375l1.828125 -1.078125l0 2.453125l1.25 0z" fill-rule="nonzero"/><path fill="#000000" d="m402.57074 241.25755l1.875 -0.171875q0.171875 0.9375 0.6875 1.390625q0.515625 0.4375 1.390625 0.4375q0.9375 0 1.40625 -0.390625q0.46875 -0.40625 0.46875 -0.921875q0 -0.34375 -0.203125 -0.578125q-0.1875 -0.234375 -0.6875 -0.421875q-0.34375 -0.109375 -1.546875 -0.40625q-1.546875 -0.390625 -2.171875 -0.953125q-0.875 -0.78125 -0.875 -1.921875q0 -0.71875 0.40625 -1.359375q0.421875 -0.640625 1.1875 -0.96875q0.78125 -0.328125 1.890625 -0.328125q1.796875 0 2.703125 0.796875q0.90625 0.78125 0.953125 2.09375l-1.921875 0.078125q-0.125 -0.734375 -0.53125 -1.046875q-0.40625 -0.328125 -1.234375 -0.328125q-0.828125 0 -1.3125 0.34375q-0.296875 0.21875 -0.296875 0.59375q0 0.34375 0.28125 0.578125q0.359375 0.3125 1.765625 0.640625q1.40625 0.328125 2.078125 0.6875q0.6875 0.359375 1.0625 0.96875q0.375 0.609375 0.375 1.515625q0 0.828125 -0.453125 1.546875q-0.453125 0.703125 -1.296875 1.0625q-0.828125 0.34375 -2.0625 0.34375q-1.8125 0 -2.78125 -0.828125q-0.96875 -0.84375 -1.15625 -2.453125zm12.5303955 -3.796875l0 1.453125l-1.25 0l0 2.78125q0 0.84375 0.03125 0.984375q0.046875 0.140625 0.171875 0.234375q0.125 0.09375 0.3125 0.09375q0.25 0 0.734375 -0.171875l0.15625 1.421875q-0.640625 0.265625 -1.453125 0.265625q-0.484375 0 -0.890625 -0.15625q-0.390625 -0.171875 -0.578125 -0.4375q-0.1875 -0.265625 -0.265625 -0.71875q-0.046875 -0.3125 -0.046875 -1.28125l0 -3.015625l-0.84375 0l0 -1.453125l0.84375 0l0 -1.375l1.828125 -1.078125l0 2.453125l1.25 0zm0.8446045 3.359375q0 -0.921875 0.453125 -1.765625q0.453125 -0.859375 1.265625 -1.296875q0.828125 -0.453125 1.84375 -0.453125q1.5625 0 2.5625 1.015625q1.015625 1.015625 1.015625 2.578125q0 1.5625 -1.015625 2.59375q-1.015625 1.03125 -2.546875 1.03125q-0.953125 0 -1.8125 -0.421875q-0.859375 -0.4375 -1.3125 -1.265625q-0.453125 -0.828125 -0.453125 -2.015625zm1.875 0.09375q0 1.03125 0.484375 1.578125q0.5 0.546875 1.203125 0.546875q0.71875 0 1.203125 -0.546875q0.484375 -0.546875 0.484375 -1.59375q0 -1.015625 -0.484375 -1.5625q-0.484375 -0.546875 -1.203125 -0.546875q-0.703125 0 -1.203125 0.546875q-0.484375 0.546875 -0.484375 1.578125zm8.438232 3.453125l-1.828125 0l0 -6.90625l1.703125 0l0 0.984375q0.4375 -0.703125 0.78125 -0.921875q0.34375 -0.21875 0.796875 -0.21875q0.625 0 1.203125 0.34375l-0.5625 1.59375q-0.46875 -0.296875 -0.859375 -0.296875q-0.390625 0 -0.65625 0.21875q-0.265625 0.203125 -0.421875 0.765625q-0.15625 0.546875 -0.15625 2.296875l0 2.140625zm7.436798 -2.203125l1.828125 0.3125q-0.34375 1.0 -1.109375 1.53125q-0.765625 0.515625 -1.890625 0.515625q-1.8125 0 -2.671875 -1.171875q-0.6875 -0.953125 -0.6875 -2.390625q0 -1.71875 0.890625 -2.6875q0.90625 -0.96875 2.28125 -0.96875q1.546875 0 2.4375 1.015625q0.890625 1.015625 0.84375 3.125l-4.578125 0q0.015625 0.8125 0.4375 1.265625q0.421875 0.453125 1.0625 0.453125q0.421875 0 0.71875 -0.234375q0.296875 -0.234375 0.4375 -0.765625zm0.109375 -1.84375q-0.015625 -0.796875 -0.40625 -1.203125q-0.390625 -0.421875 -0.953125 -0.421875q-0.59375 0 -0.984375 0.4375q-0.390625 0.4375 -0.390625 1.1875l2.734375 0z" fill-rule="nonzero"/><path fill="#000000" fill-opacity="0.0" d="m315.073 197.48778l-34.236237 0.06298828" fill-rule="evenodd"/><path stroke="#000000" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m311.283 197.49475l-26.65625 0.049041748" fill-rule="evenodd"/><path fill="#000000" stroke="#000000" stroke-width="1.0" stroke-linecap="butt" d="m314.57297 197.4887c0.0016784668 0.9085083 -0.7334595 1.6463623 -1.6419678 1.6480255c-0.9085083 0.0016784668 -1.6463318 -0.7334595 -1.6480103 -1.6419678c-0.0016784668 -0.9085083 0.7334595 -1.646347 1.6419678 -1.6480255c0.9085083 -0.001663208 1.6463623 0.7334595 1.6480103 1.6419678z" fill-rule="nonzero"/><path fill="#000000" stroke="#000000" stroke-width="1.0" stroke-linecap="butt" d="m281.33676 197.54985c-0.0016784668 -0.9085083 0.7334595 -1.646347 1.6419678 -1.6480255c0.9085083 -0.001663208 1.6463623 0.7334595 1.6480103 1.6419678c0.0016784668 0.9085083 -0.7334595 1.6463623 -1.6419678 1.6480255c-0.9085083 0.0016784668 -1.6463318 -0.7334595 -1.6480103 -1.6419678z" fill-rule="nonzero"/><g filter="url(#shadowFilter-p.3)"><use xlink:href="#p.3" transform="matrix(1.0 0.0 0.0 1.0 0.0 2.0)"/></g><defs><filter id="shadowFilter-p.3" filterUnits="userSpaceOnUse"><feGaussianBlur in="SourceAlpha" stdDeviation="2.0" result="blur"/><feComponentTransfer in="blur" color-interpolation-filters="sRGB"><feFuncR type="linear" slope="0" intercept="0.0"/><feFuncG type="linear" slope="0" intercept="0.0"/><feFuncB type="linear" slope="0" intercept="0.0"/><feFuncA type="linear" slope="0.5" intercept="0"/></feComponentTransfer></filter></defs><g id="p.3"><path fill="#7400ff" d="m-0.23622048 181.01312l119.433075 0l0 34.48819l-119.433075 0z" fill-rule="evenodd"/><path stroke="#ffffff" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m-0.23622048 181.01312l119.433075 0l0 34.48819l-119.433075 0z" fill-rule="evenodd"/><path fill="#ffffff" d="m29.064884 200.7591l1.59375 -7.84375l1.796875 0l-2.40625 10.421875l-1.71875 0l-1.984375 -7.609375l-2.03125 7.609375l-1.734375 0l-2.40625 -10.421875l1.796875 0l1.609375 7.828125l2.0 -7.828125l1.515625 0l1.96875 7.84375zm4.0019073 -1.359375q0 -1.140625 0.453125 -2.046875q0.453125 -0.921875 1.265625 -1.40625q0.8125 -0.5 1.875 -0.5q1.5625 0 2.53125 1.015625q0.984375 1.0 1.0625 2.671875l0.015625 0.40625q0 1.15625 -0.4375 2.0625q-0.4375 0.890625 -1.265625 1.390625q-0.8125 0.484375 -1.890625 0.484375q-1.640625 0 -2.625 -1.09375q-0.984375 -1.09375 -0.984375 -2.90625l0 -0.078125zm1.734375 0.140625q0 1.203125 0.5 1.875q0.5 0.671875 1.375 0.671875q0.875 0 1.359375 -0.6875q0.5 -0.6875 0.5 -2.0q0 -1.171875 -0.5 -1.859375q-0.5 -0.6875 -1.375 -0.6875q-0.84375 0 -1.359375 0.671875q-0.5 0.671875 -0.5 2.015625zm10.984955 -2.359375q-0.34375 -0.0625 -0.703125 -0.0625q-1.1875 0 -1.609375 0.921875l0 5.296875l-1.734375 0l0 -7.75l1.65625 0l0.046875 0.875q0.625 -1.015625 1.75 -1.015625q0.359375 0 0.609375 0.109375l-0.015625 1.625zm3.621338 2.828125l-0.765625 0.796875l0 2.53125l-1.75 0l0 -11.0l1.75 0l0 6.34375l0.53125 -0.671875l2.140625 -2.421875l2.09375 0l-2.875 3.234375l3.1875 4.515625l-2.015625 0l-2.296875 -3.328125zm9.478287 1.21875q0 -0.453125 -0.390625 -0.703125q-0.375 -0.25 -1.265625 -0.421875q-0.890625 -0.1875 -1.484375 -0.484375q-1.296875 -0.625 -1.296875 -1.8125q0 -1.015625 0.84375 -1.6875q0.84375 -0.671875 2.140625 -0.671875q1.390625 0 2.25 0.6875q0.859375 0.6875 0.859375 1.78125l-1.75 0q0 -0.5 -0.375 -0.828125q-0.359375 -0.328125 -0.984375 -0.328125q-0.5625 0 -0.9375 0.265625q-0.359375 0.265625 -0.359375 0.703125q0 0.40625 0.328125 0.625q0.34375 0.21875 1.359375 0.453125q1.03125 0.21875 1.609375 0.53125q0.59375 0.3125 0.875 0.75q0.28125 0.4375 0.28125 1.046875q0 1.046875 -0.875 1.703125q-0.859375 0.640625 -2.25 0.640625q-0.953125 0 -1.703125 -0.34375q-0.75 -0.34375 -1.171875 -0.9375q-0.40625 -0.609375 -0.40625 -1.296875l1.6875 0q0.046875 0.609375 0.46875 0.953125q0.4375 0.328125 1.140625 0.328125q0.6875 0 1.046875 -0.265625q0.359375 -0.265625 0.359375 -0.6875zm5.3924065 -7.515625l0 1.875l1.375 0l0 1.296875l-1.375 0l0 4.328125q0 0.4375 0.171875 0.640625q0.1875 0.1875 0.640625 0.1875q0.296875 0 0.609375 -0.0625l0 1.34375q-0.59375 0.15625 -1.15625 0.15625q-1.9999962 0 -1.9999962 -2.203125l0 -4.390625l-1.265625 0l0 -1.296875l1.265625 0l0 -1.875l1.7343712 0zm7.3266144 9.625q-0.109375 -0.21875 -0.203125 -0.71875q-0.828125 0.859375 -2.03125 0.859375q-1.15625 0 -1.90625 -0.65625q-0.734375 -0.671875 -0.734375 -1.65625q0 -1.234375 0.921875 -1.890625q0.921875 -0.671875 2.625 -0.671875l1.078125 0l0 -0.5q0 -0.609375 -0.34375 -0.96875q-0.328125 -0.359375 -1.015625 -0.359375q-0.59375 0 -0.984375 0.296875q-0.375 0.296875 -0.375 0.75l-1.734375 0q0 -0.625 0.421875 -1.171875q0.421875 -0.5625 1.140625 -0.875q0.734375 -0.328125 1.625 -0.328125q1.359375 0 2.171875 0.6875q0.8125 0.6875 0.828125 1.921875l0 3.484375q0 1.046875 0.296875 1.671875l0 0.125l-1.78125 0zm-1.90625 -1.25q0.515625 0 0.96875 -0.25q0.453125 -0.25 0.6875 -0.671875l0 -1.46875l-0.9375 0q-0.96875 0 -1.453125 0.34375q-0.484375 0.328125 -0.484375 0.953125q0 0.5 0.328125 0.796875q0.328125 0.296875 0.890625 0.296875zm7.3823853 -8.375l0 1.875l1.375 0l0 1.296875l-1.375 0l0 4.328125q0 0.4375 0.171875 0.640625q0.1875 0.1875 0.640625 0.1875q0.296875 0 0.609375 -0.0625l0 1.34375q-0.59375 0.15625 -1.15625 0.15625q-2.0 0 -2.0 -2.203125l0 -4.390625l-1.265625 0l0 -1.296875l1.265625 0l0 -1.875l1.734375 0zm4.5453644 9.625l-1.734375 0l0 -7.75l1.734375 0l0 7.75zm-1.84375 -9.75q0 -0.40625 0.25 -0.671875q0.265625 -0.265625 0.734375 -0.265625q0.46875 0 0.71875 0.265625q0.265625 0.265625 0.265625 0.671875q0 0.390625 -0.265625 0.65625q-0.25 0.25 -0.71875 0.25q-0.46875 0 -0.734375 -0.25q-0.25 -0.265625 -0.25 -0.65625zm3.4146576 5.8125q0 -1.140625 0.453125 -2.046875q0.453125 -0.921875 1.265625 -1.40625q0.8125 -0.5 1.875 -0.5q1.5625 0 2.53125 1.015625q0.984375 1.0 1.0625 2.671875l0.015625 0.40625q0 1.15625 -0.4375 2.0625q-0.4375 0.890625 -1.265625 1.390625q-0.8125 0.484375 -1.890625 0.484375q-1.640625 0 -2.625 -1.09375q-0.984375 -1.09375 -0.984375 -2.90625l0 -0.078125zm1.734375 0.140625q0 1.203125 0.5 1.875q0.5 0.671875 1.375 0.671875q0.875 0 1.359375 -0.6875q0.5 -0.6875 0.5 -2.0q0 -1.171875 -0.5 -1.859375q-0.5 -0.6875 -1.375 -0.6875q-0.84375 0 -1.359375 0.671875q-0.5 0.671875 -0.5 2.015625zm8.547455 -3.953125l0.0625 0.90625q0.84375 -1.046875 2.25 -1.046875q2.421875 0 2.453125 2.78125l0 5.109375l-1.734375 0l0 -5.015625q0 -0.734375 -0.328125 -1.09375q-0.3125 -0.359375 -1.03125 -0.359375q-1.046875 0 -1.5625 0.953125l0 5.515625l-1.75 0l0 -7.75l1.640625 0z" fill-rule="nonzero"/></g><path fill="#000000" fill-opacity="0.0" d="m119.19685 198.25722l145.03935 -0.69291687" fill-rule="evenodd"/><path stroke="#000000" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m122.9868 198.2391l137.45946 -0.6566925" fill-rule="evenodd"/><path fill="#000000" stroke="#000000" stroke-width="1.0" stroke-linecap="butt" d="m119.696846 198.25482c-0.0043411255 -0.90849304 0.72862244 -1.6484985 1.6371231 -1.652832c0.90849304 -0.004333496 1.6484985 0.72862244 1.6528397 1.6371155c0.0043411255 0.9085083 -0.72862244 1.6484985 -1.6371231 1.6528473c-0.9085007 0.004333496 -1.6484985 -0.72862244 -1.6528397 -1.6371307z" fill-rule="nonzero"/><path fill="#000000" stroke="#000000" stroke-width="1.0" stroke-linecap="butt" d="m263.73624 197.5667c0.004333496 0.90849304 -0.7286377 1.6484985 -1.637146 1.652832c-0.9084778 0.004348755 -1.6484985 -0.72862244 -1.652832 -1.6371155c-0.004333496 -0.90849304 0.7286377 -1.6484985 1.6371155 -1.6528473c0.9085083 -0.004333496 1.6484985 0.72862244 1.6528625 1.6371307z" fill-rule="nonzero"/><path fill="#4a86e8" d="m54.761154 118.40164l0 0c0 -1.820694 1.4759636 -3.2966537 3.2966537 -3.2966537l154.66653 0c0.87431335 0 1.7128448 0.34732056 2.3310852 0.96556854c0.61824036 0.61824036 0.9655609 1.4567566 0.9655609 2.3310852l0 13.186218c0 1.820694 -1.4759674 3.2966614 -3.296646 3.2966614l-154.66653 0c-1.8206902 0 -3.2966537 -1.4759674 -3.2966537 -3.2966614z" fill-rule="evenodd"/><path stroke="#000000" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m54.761154 118.40164l0 0c0 -1.820694 1.4759636 -3.2966537 3.2966537 -3.2966537l154.66653 0c0.87431335 0 1.7128448 0.34732056 2.3310852 0.96556854c0.61824036 0.61824036 0.9655609 1.4567566 0.9655609 2.3310852l0 13.186218c0 1.820694 -1.4759674 3.2966614 -3.296646 3.2966614l-154.66653 0c-1.8206902 0 -3.2966537 -1.4759674 -3.2966537 -3.2966614z" fill-rule="evenodd"/><path fill="#ffffff" d="m69.54472 128.73476l0 -7.6250076l2.3125 0l1.375 5.203125l1.375 -5.203125l2.3125 0l0 7.6250076l-1.4375 0l0 -6.0000076l-1.515625 6.0000076l-1.484375 0l-1.5 -6.0000076l0 6.0000076l-1.4375 0zm8.9392395 0l0 -7.5625076l1.546875 0l0 6.28125l3.828125 0l0 1.2812576l-5.375 0zm6.478012 0l0 -7.6250076l5.234375 0l0 1.28125l-3.6875 0l0 1.8125l3.1875 0l0 1.296875l-3.1875 0l0 3.2343826l-1.546875 0zm6.540512 0l0 -7.5625076l1.546875 0l0 6.28125l3.828125 0l0 1.2812576l-5.375 0zm6.165512 -3.7656326q0 -1.171875 0.34375 -1.953125q0.265625 -0.59375 0.703125 -1.046875q0.453125 -0.46875 1.0 -0.6875q0.703125 -0.3125 1.640625 -0.3125q1.6875 0 2.6875 1.046875q1.015625 1.046875 1.015625 2.921875q0 1.84375 -1.0 2.890625q-1.0 1.0312576 -2.6875 1.0312576q-1.703125 0 -2.703125 -1.0312576q-1.0 -1.046875 -1.0 -2.859375zm1.578125 -0.046875q0 1.28125 0.59375 1.953125q0.609375 0.671875 1.53125 0.671875q0.921875 0 1.5 -0.65625q0.59375 -0.671875 0.59375 -2.0q0 -1.3125 -0.578125 -1.953125q-0.5625 -0.640625 -1.515625 -0.640625q-0.953125 0 -1.546875 0.65625q-0.578125 0.640625 -0.578125 1.96875zm8.101273 3.8125076l-1.828125 -7.6250076l1.578125 0l1.15625 5.234375l1.390625 -5.234375l1.828125 0l1.34375 5.328125l1.171875 -5.328125l1.546875 0l-1.859375 7.6250076l-1.625 0l-1.515625 -5.7031326l-1.515625 5.7031326l-1.671875 0zm8.104752 2.109375l0 -0.953125l6.078125 0l0 0.953125l-6.078125 0zm8.520248 -2.109375l0 -6.3437576l-2.265625 0l0 -1.28125l6.0625 0l0 1.28125l-2.265625 0l0 6.3437576l-1.53125 0zm4.790512 0l0 -7.6250076l3.234375 0q1.234375 0 1.78125 0.203125q0.5625 0.203125 0.890625 0.734375q0.328125 0.515625 0.328125 1.203125q0 0.859375 -0.5 1.421875q-0.5 0.546875 -1.515625 0.703125q0.5 0.28125 0.828125 0.640625q0.328125 0.34375 0.875 1.234375l0.9375 1.4843826l-1.84375 0l-1.109375 -1.6562576q-0.59375 -0.890625 -0.8125 -1.125q-0.21875 -0.234375 -0.46875 -0.3125q-0.25 -0.09375 -0.765625 -0.09375l-0.3125 0l0 3.1875076l-1.546875 0zm1.546875 -4.4062576l1.125 0q1.109375 0 1.390625 -0.09375q0.28125 -0.09375 0.4375 -0.3125q0.15625 -0.234375 0.15625 -0.578125q0 -0.390625 -0.21875 -0.625q-0.203125 -0.234375 -0.578125 -0.296875q-0.1875 -0.03125 -1.125 -0.03125l-1.1875 0l0 1.9375zm13.023727 4.4062576l-1.671875 0l-0.671875 -1.7343826l-3.046875 0l-0.625 1.7343826l-1.640625 0l2.96875 -7.6250076l1.625 0l3.0625 7.6250076zm-2.84375 -3.0156326l-1.046875 -2.828125l-1.03125 2.828125l2.078125 0zm8.539352 0.21875l1.5 0.46875q-0.34375 1.25 -1.15625 1.8593826q-0.796875 0.59375 -2.015625 0.59375q-1.515625 0 -2.5 -1.0312576q-0.984375 -1.046875 -0.984375 -2.84375q0 -1.90625 0.984375 -2.953125q0.984375 -1.0625 2.59375 -1.0625q1.390625 0 2.265625 0.828125q0.53125 0.5 0.78125 1.40625l-1.515625 0.375q-0.140625 -0.59375 -0.578125 -0.9375q-0.421875 -0.34375 -1.03125 -0.34375q-0.859375 0 -1.390625 0.609375q-0.515625 0.609375 -0.515625 1.96875q0 1.4375 0.515625 2.0625q0.515625 0.609375 1.359375 0.609375q0.609375 0 1.046875 -0.390625q0.453125 -0.390625 0.640625 -1.21875zm2.8362274 2.7968826l0 -7.6250076l1.546875 0l0 3.390625l3.109375 -3.390625l2.0625 0l-2.875 2.96875l3.03125 4.6562576l-1.984375 0l-2.109375 -3.5781326l-1.234375 1.265625l0 2.3125076l-1.546875 0zm7.6331024 0l0 -7.6250076l1.53125 0l0 7.6250076l-1.53125 0zm3.0231476 0l0 -7.6250076l1.5 0l3.109375 5.09375l0 -5.09375l1.4375 0l0 7.6250076l-1.546875 0l-3.078125 -4.9687576l0 4.9687576l-1.421875 0zm11.226852 -2.7968826l0 -1.296875l3.3125 0l0 3.046875q-0.484375 0.46875763 -1.40625 0.8281326q-0.90625 0.34375 -1.859375 0.34375q-1.1875 0 -2.078125 -0.5q-0.890625 -0.5000076 -1.34375 -1.4218826q-0.4375 -0.9375 -0.4375 -2.046875q0 -1.1875 0.5 -2.109375q0.5 -0.9375 1.453125 -1.421875q0.734375 -0.390625 1.828125 -0.390625q1.421875 0 2.21875 0.609375q0.796875 0.59375 1.03125 1.640625l-1.53125 0.28125q-0.15625 -0.5625 -0.609375 -0.875q-0.4375 -0.328125 -1.109375 -0.328125q-1.015625 0 -1.609375 0.640625q-0.59375 0.640625 -0.59375 1.890625q0 1.359375 0.59375 2.046875q0.609375 0.671875 1.59375 0.671875q0.484375 0 0.96875 -0.1875q0.484375 -0.1875 0.84375 -0.46875l0 -0.953125l-1.765625 0zm3.8668976 4.9062576l0 -0.953125l6.078125 0l0 0.953125l-6.078125 0zm6.7858734 -9.734383l1.546875 0l0 4.125q0 0.984375 0.046875 1.28125q0.109375 0.46875 0.46875 0.75q0.375 0.28125 1.015625 0.28125q0.65625 0 0.984375 -0.265625q0.34375 -0.265625 0.40625 -0.65625q0.078125 -0.390625 0.078125 -1.296875l0 -4.21875l1.53125 0l0 4.0q0 1.375 -0.125 1.953125q-0.125 0.5625 -0.46875 0.9531326q-0.328125 0.390625 -0.890625 0.625q-0.5625 0.21875 -1.46875 0.21875q-1.09375 0 -1.65625 -0.25q-0.5625 -0.25 -0.890625 -0.6562576q-0.328125 -0.40625 -0.421875 -0.84375q-0.15625 -0.65625 -0.15625 -1.9375l0 -4.0625zm7.7112274 7.6250076l0 -7.6250076l3.234375 0q1.234375 0 1.78125 0.203125q0.5625 0.203125 0.890625 0.734375q0.328125 0.515625 0.328125 1.203125q0 0.859375 -0.5 1.421875q-0.5 0.546875 -1.515625 0.703125q0.5 0.28125 0.828125 0.640625q0.328125 0.34375 0.875 1.234375l0.9375 1.4843826l-1.84375 0l-1.109375 -1.6562576q-0.59375 -0.890625 -0.8125 -1.125q-0.21875 -0.234375 -0.46875 -0.3125q-0.25 -0.09375 -0.765625 -0.09375l-0.3125 0l0 3.1875076l-1.546875 0zm1.546875 -4.4062576l1.125 0q1.109375 0 1.390625 -0.09375q0.28125 -0.09375 0.4375 -0.3125q0.15625 -0.234375 0.15625 -0.578125q0 -0.390625 -0.21875 -0.625q-0.203125 -0.234375 -0.578125 -0.296875q-0.1875 -0.03125 -1.125 -0.03125l-1.1875 0l0 1.9375zm6.1018524 4.4062576l0 -7.6250076l1.53125 0l0 7.6250076l-1.53125 0z" fill-rule="nonzero"/><path fill="#000000" fill-opacity="0.0" d="m135.39108 134.88452l0 25.165344l128.8819 0" fill-rule="evenodd"/><path stroke="#000000" stroke-width="1.0" stroke-linejoin="round" stroke-linecap="butt" d="m135.39108 134.88452l0 25.165344l125.4548 0" fill-rule="evenodd"/><path fill="#000000" stroke="#000000" stroke-width="1.0" stroke-linecap="butt" d="m260.8459 160.04987l-1.1246033 1.124588l3.0897827 -1.124588l-3.0897827 -1.124588z" fill-rule="evenodd"/></g></svg> | 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/src | rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/src/rf_test/train.py | """Hyperparameter optimization with cuML, hyperopt, and MLFlow"""
import argparse
from functools import partial
import sys
import gcsfs
import mlflow
import mlflow.sklearn
import cuml
import cudf
from cuml.metrics.accuracy import accuracy_score
from cuml.model_selection import train_test_split
from cuml.ensemble import RandomForestClassifier
from hyperopt import fmin, tpe, hp, Trials, STATUS_OK
import traceback
def load_data(fpath):
"""
Simple helper function for loading data to be used by CPU/GPU models.
:param fpath: Path to the data to be ingested
:return: DataFrame wrapping the data at [fpath]. Data will be in either a Pandas or RAPIDS (cuDF) DataFrame
"""
import cudf
if (fpath.startswith('gs://')):
fs = gcsfs.GCSFileSystem()
with fs.open(fpath, mode='rb') as f:
df = cudf.read_parquet(f)
else:
df = cudf.read_parquet(fpath)
X = df.drop(["ArrDelayBinary"], axis=1)
y = df["ArrDelayBinary"].astype("int32")
return train_test_split(X, y, test_size=0.2)
def _train(params, fpath, hyperopt=False):
"""
:param params: hyperparameters. Its structure is consistent with how search space is defined. See below.
:param fpath: Path or URL for the training data used with the model.
:param hyperopt: Use hyperopt for hyperparameter search during training.
:return: dict with fields 'loss' (scalar loss) and 'status' (success/failure status of run)
"""
max_depth, max_features, n_estimators = params
max_depth, max_features, n_estimators = (int(max_depth), float(max_features), int(n_estimators))
# Log all of our training parameters for this run.
pyver = sys.version_info
mlparams = {
'cudf_version': str(cudf.__version__),
'cuml_version': str(cuml.__version__),
'max_depth': str(max_depth),
'max_features': str(max_features),
'n_estimators': str(n_estimators),
'python_version': f"{pyver[0]}.{pyver[1]}.{pyver[2]}.{pyver[3]}",
}
mlflow.log_params(mlparams)
X_train, X_test, y_train, y_test = load_data(fpath)
mod = RandomForestClassifier(
max_depth=max_depth, max_features=max_features, n_estimators=n_estimators
)
mod.fit(X_train, y_train)
preds = mod.predict(X_test)
acc = accuracy_score(y_test, preds)
mlflow.log_metric("accuracy", acc)
mlflow.sklearn.log_model(mod, "saved_models")
if not hyperopt:
return mod
return {"loss": acc, "status": STATUS_OK}
def train(params, fpath, hyperopt=False):
"""
Proxy function used to call _train
:param params: hyperparameters. Its structure is consistent with how search space is defined. See below.
:param fpath: Path or URL for the training data used with the model.
:param hyperopt: Use hyperopt for hyperparameter search during training.
:return: dict with fields 'loss' (scalar loss) and 'status' (success/failure status of run)
"""
with mlflow.start_run(nested=True):
return _train(params, fpath, hyperopt)
def prep_env(args):
cpath = args.conda_env
if (cpath.startswith('gs://')):
fs = gcsfs.GCSFileSystem()
with fs.open(cpath, mode='r') as f:
cfile = f.read()
with open('envs/conda.yaml', 'w') as writer:
writer.write(cfile)
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument("--algo", default="tpe", choices=["tpe"], type=str)
parser.add_argument("--conda-env", required=True, type=str)
parser.add_argument("--fpath", required=True, type=str)
args = parser.parse_args()
search_space = [
hp.uniform("max_depth", 5, 20),
hp.uniform("max_features", 0.1, 1.0),
hp.uniform("n_estimators", 150, 1000),
]
prep_env(args)
try:
trials = Trials()
algorithm = tpe.suggest if args.algo == "tpe" else None
fn = partial(train, fpath=args.fpath, hyperopt=True)
artifact_path = "airline-demo"
with mlflow.start_run(run_name="RAPIDS-Hyperopt"):
argmin = fmin(fn=fn, space=search_space, algo=algorithm, max_evals=10, trials=trials)
print("===========")
fn = partial(train, fpath=args.fpath, hyperopt=False)
final_model = fn(tuple(argmin.values()))
mlflow.sklearn.log_model(
final_model,
artifact_path=artifact_path,
registered_model_name="rapids_airline_hyperopt_k8s",
conda_env="envs/conda.yaml",
)
except Exception as e:
print(f"train.py threw exception: {e}", flush=True)
traceback.print_exc()
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/src | rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/src/rf_test/test_query.py | import os
import json
import requests
host = 'localhost'
port = '56767'
headers = {
"Content-Type": "application/json",
"format": "pandas-split"
}
data = {
"columns": ["Year", "Month", "DayofMonth", "DayofWeek", "CRSDepTime", "CRSArrTime", "UniqueCarrier",
"FlightNum", "ActualElapsedTime", "Origin", "Dest", "Distance", "Diverted"],
"data": [[1987, 10, 1, 4, 1, 556, 0, 190, 247, 202, 162, 1846, 0]]
}
resp = requests.post(url="http://%s:%s/invocations" % (host, port), data=json.dumps(data), headers=headers)
print('Classification: %s' % ("ON-Time" if resp.text == "[0.0]" else "LATE"))
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/src | rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/src/k8s/entrypoint.sh | #!/bin/bash
# Activates the correct Anaconda environment, and runs the command passed to the container.
set -e
set -x
source activate rapids
nvidia-smi
ARGS=( "$@" )
python --version
echo "Calling: 'python ${ARGS[@]}'"
python ${ARGS[@]}
echo "Python call returned: $?"
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/src | rapidsai_public_repos/cloud-ml-examples/mlflow/docker_environment/src/k8s/tracking_entrypoint.sh | #/bin/bash
# Launch our mlflow tracking server
set -e
set -x
mlflow server --backend-store-uri=postgresql://${MLFLOW_USER}:${MLFLOW_PASS}@${MLFLOW_DB_ADDR}:${MLFLOW_DB_PORT}/${MLFLOW_DB_NAME} --default-artifact-root=${MLFLOW_ARTIFACT_PATH} --host 0.0.0.0 --port 80
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow | rapidsai_public_repos/cloud-ml-examples/mlflow/local_environment/README-Databricks.md | ### Utilizing Databricks' for MLFlow Tracking and Job Training
#### Assumptions and Naming Conventions
- All shell commands are assumed to be run within the `/cloud-ml-examples/mlflow/docker_environment` directory.
- There are a number of configuration parameters that will be specific to your _environment_ and _deployment_:
- `DATBRICKS HOST` : URI of your Databricks service, will be of the form: `https://<cluster_id>.cloud.databricks.com`
- `DATABRICKS TOKEN` : Access token used to authenticate with your Databricks service account
- [Token Creation Process](https://docs.databricks.com/dev-tools/api/latest/authentication.html#:~:text=Generate%20a%20personal%20access%20token,-This%20section%20describes&text=in%20the%20upper%20right%20corner,the%20Generate%20New%20Token%20button.).
See the [Databricks documentation](https://docs.databricks.com/applications/mlflow/access-hosted-tracking-server.html) for additional information.
- `EXPERIMENT NAME` : MLflow experiment name, which will be used to register with the tracking server and indicates
the subdirectory of your user environment to write to.
- `YOUR USER NAME` : Databricks username
#### Databricks-Requirements
1. Set environment variables expected by MLFlow, which indicate the location of your Databricks cluster and how to connect to it.
- **Note**: If you make any changes to the project repo, they will not be reflected in your mlflow deployment until
they are committed to the working branch.
- **Note**: While the MLFLow Python API interface should allow these values to be set programatically, that workflow
does not appear to work; as of version 1.8.0, `set_tracking_uri` and `set_experiment` do not produce the expected behavior when targeting Databricks. You will likely see raised errors claiming that active and selected experiments do not match, and/or spurrious authentication errors.
```bash
$ export MLFLOW_EXPERIMENT_NAME=/Users/[YOUR USER NAME]/[EXPERIMENT NAME]
$ export MLFLOW_TRACKING_URI=databricks
$ export DATABRICKS_HOST=https://[CLUSTER_ID].cloud.databricks.com
$ export DATABRICKS_TOKEN="[ACCESS TOKEN]"
```
#### Train Locally and Publish to Databrick's MLFlow Tracking Server.
[Client and Tracking APIs](https://www.mlflow.org/docs/latest/tracking.html).
- Create a new conda environment, and configure your [databricks-cli](https://docs.databricks.com/dev-tools/cli/index.html) client.
- `$ conda create -f envs/conda.yaml`
- `$ databricks configure`
- Train the model
- `$ cd mlflow`
- Publish to Databrick's tracking server
- Here, we use mlflow to run our training routine locally, publish the resulting metrics to our configured
Databricks account, and save our RAPIDS model.
- Export the [required environment](#databricks-requirements) variables for MLFlow
```shell script
mlflow run file:///$PWD -b local -e hyperopt \
-P conda-env=$PWD/envs/conda.yaml\
-P fpath=https://rapidsai-cloud-ml-sample-data.s3-us-west-2.amazonaws.com/airline_small.parquet
```
- Deploy your model
- Locate the model's 'Full Path' identity.
- Databricks
- Locate your saved experiment in the Databricks tracking UI at: `/Users/[YOUR USER NAME]/[EXPERIMENT NAME]`
- Ex. `dbfs:/databricks/mlflow/[EXPERIMENT ID #]/[EXPERIMENT RUN HASH]/artifacts/`
- Select the successful run and find the 'Full Path' element
1. 
- Deploy your model
- If you have not defined the environment variables described above, this will fail to pull your model
from Databricks.
- `mlflow models serve -m [PATH_TO_MODEL] -p 55755`
- From Databricks MLFlow UI
- Locate your saved experiment in the Databricks tracking UI at: `/Users/[YOUR USER NAME]/[EXPERIMENT NAME]`
- Ex. `dbfs:/databricks/mlflow/[EXPERIMENT ID #]/[EXPERIMENT RUN HASH]/artifacts/`
- `mlflow models serve -m [PATH_TO_MODEL] -p 55755`
- Query the deployed model with test data `test_call.sh` example script.
- `bash test_call.sh`
## Train Models Using MLFlow with Hyperopt and RAPIDS on Databricks.
#### Currently, this approach does not support SparkTrials integration with Hyperopt.
- Define a cluster configuration, or use the [sample provided](cluster_definitions/training_cluster.json)
- Initiate Databricks job
```shell script
mlflow run file:///$PWD -b databricks \
--backend-config=./databricks/training_cluster.json \
-P conda-env=[CONDA SPEC PATH OR URL] \
-P fpath=[DATH PATH]
```
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow | rapidsai_public_repos/cloud-ml-examples/mlflow/local_environment/README.md | ### Train and Publish Locally With MLFlow
#### Jupyter Notebook Workflow
[Jupyter Notebook](notebooks/rapids_mlflow_databricks_train_deploy.ipynb)
#### To reproduce this workflow, utilizing Databricks MLFlow tracking server, see:
- [Databricks MLFlow CLI](README-Databricks.md)
#### CLI Based Workflow
- Create a new conda environment.
- `$ conda create -f envs/conda.yaml`
- Train the model
- `$ cd mlflow`
- MLflow project configuration is described in our [MLProject](https://www.mlflow.org/docs/latest/projects.html) file.
- This can be edited to allow additional command line variables, specify conda environments, and training
parameters (see link for additional information).
- Publish to local tracking server
- Here use mlflow to run our training routine locally, and publish the results to the local file system.
- In your shell, run:
```shell script
# Downlad the file
wget -N https://rapidsai-cloud-ml-sample-data.s3-us-west-2.amazonaws.com/airline_small.parquet
# Launch the job
mlflow run . -b local -e hyperopt \
-P conda-env=$PWD/envs/conda.yaml \
-P fpath=airline_small.parquet
```
- Deploy your model
- Locate the model's 'Full Path'
- `mlflow ui`
- Locate the model path using the mlflow ui at localhost:5000
- Select the successful run and find the 'Full Path' element
- 
- Deploy your model
- `$ mlflow models serve -m [PATH_TO_MODEL] -p 55755`
1. Query the deployed model with test data `src/sample_server_query.sh` example script.
1. `bash src/sample_server_query.sh`
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow | rapidsai_public_repos/cloud-ml-examples/mlflow/local_environment/MLproject | name: cuML RF test
conda_env: envs/conda.yaml
entry_points:
hyperopt:
parameters:
fpath: {type: str}
algo: {type: str, default: 'tpe'}
command: "python src/rf_test/train.py --fpath={fpath} --algo={algo}"
simple:
parameters:
conda_env: {type: str, default: './envs/conda.yaml'}
fpath: {type: str, default: './airline_small.parquet'}
n_estimators: {type: int, default: 100}
max_features: {type: float, default: 0.33}
max_depth: {type: int, default: 10}
command: "python src/rf_test/train_simple.py --fpath={fpath} --n_estimators={n_estimators} \
--max_features={max_features} --max_depth={max_depth} --conda_env={conda_env}"
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/local_environment | rapidsai_public_repos/cloud-ml-examples/mlflow/local_environment/cluster_definitions/training_cluster.json | {
"autoscale": {
"min_workers": 7,
"max_workers": 8
},
"spark_version": "6.6.x-gpu-ml-scala2.11",
"spark_conf": {
"spark.executor.cores": "2"
},
"aws_attributes": {
"first_on_demand": 1,
"availability": "SPOT_WITH_FALLBACK",
"zone_id": "us-west-2a",
"spot_bid_price_percent": 100,
"ebs_volume_type": "GENERAL_PURPOSE_SSD",
"ebs_volume_count": 3,
"ebs_volume_size": 100
},
"node_type_id": "p3.2xlarge",
"driver_node_type_id": "p3.2xlarge",
"ssh_public_keys": [],
"custom_tags": {},
"cluster_log_conf": {
"dbfs": {
"destination": "dbfs:/cluster-logs"
}
},
"spark_env_vars": {},
"enable_elastic_disk": false,
"cluster_source": "UI",
"init_scripts": []
}
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/local_environment | rapidsai_public_repos/cloud-ml-examples/mlflow/local_environment/envs/conda.yaml | name: mlflow
channels:
- rapidsai
- nvidia
- conda-forge
dependencies:
- _libgcc_mutex=0.1=conda_forge
- _openmp_mutex=4.5=1_gnu
- abseil-cpp=20200225.2=he1b5a44_2
- appdirs=1.4.3=py_1
- arrow-cpp=0.17.1=py38h1234567_11_cuda
- arrow-cpp-proc=1.0.1=cuda
- asn1crypto=1.4.0=pyh9f0ad1d_0
- aws-c-common=0.4.57=he1b5a44_1
- aws-c-event-stream=0.1.6=h72b8ae1_3
- aws-checksums=0.1.9=h346380f_0
- aws-sdk-cpp=1.7.164=h69f4914_4
- bokeh=2.2.1=py38_0
- boost-cpp=1.72.0=h9359b55_3
- brotli=1.0.9=he6710b0_0
- brotlipy=0.7.0=py38h1e0a361_1000
- bzip2=1.0.8=h7b6447c_0
- c-ares=1.16.1=h7b6447c_0
- ca-certificates=2020.7.22=0
- certifi=2020.6.20=py38_0
- cffi=1.14.3=py38he30daa8_0
- chardet=3.0.4=py38h32f6830_1007
- click=7.1.2=py_0
- cloudpickle=1.6.0=py_0
- configparser=5.0.0=py_0
- cryptography=3.1.1=py38h766eaa4_0
- cudatoolkit=11.0.221=h6bb024c_0
- cudf=0.15.0=cuda_11.0_py38_g71cb8c0e0_0
- cudnn=8.0.0=cuda11.0_0
- cuml=0.15.0=cuda11.0_py38_ga3002e587_0
- cupy=7.8.0=py38hb7c6141_0
- cytoolz=0.11.0=py38h7b6447c_0
- dask=2.27.0=py_0
- dask-core=2.27.0=py_0
- dask-cudf=0.15.0=py38_g71cb8c0e0_0
- databricks-cli=0.9.1=py_0
- distributed=2.27.0=py38_0
- dlpack=0.3=he6710b0_1
- docker-py=4.3.1=py38h32f6830_0
- docker-pycreds=0.4.0=py_0
- double-conversion=3.1.5=he6710b0_1
- entrypoints=0.3=py38h32f6830_1001
- faiss-proc=1.0.0=cuda
- fastavro=1.0.0.post1=py38h7b6447c_0
- fastrlock=0.5=py38he6710b0_0
- flask=1.1.2=pyh9f0ad1d_0
- freetype=2.10.2=h5ab3b9f_0
- fsspec=0.8.0=py_0
- gflags=2.2.2=he6710b0_0
- gitdb=4.0.5=py_0
- gitpython=3.1.8=py_0
- glog=0.4.0=he6710b0_0
- gorilla=0.3.0=py_0
- grpc-cpp=1.30.2=heedbac9_0
- gunicorn=20.0.4=py38h32f6830_1
- heapdict=1.0.1=py_0
- icu=67.1=he1b5a44_0
- idna=2.10=pyh9f0ad1d_0
- itsdangerous=1.1.0=py_0
- jinja2=2.11.2=py_0
- joblib=0.16.0=py_0
- jpeg=9b=h024ee3a_2
- krb5=1.18.2=h173b8e3_0
- lcms2=2.11=h396b838_0
- ld_impl_linux-64=2.33.1=h53a641e_7
- libblas=3.8.0=17_openblas
- libcblas=3.8.0=17_openblas
- libcudf=0.15.0=cuda11.0_g71cb8c0e0_0
- libcuml=0.15.0=cuda11.0_ga3002e587_0
- libcumlprims=0.15.0=cuda11.0_gdbd0d39_0
- libcurl=7.71.1=h20c2e04_1
- libedit=3.1.20191231=h14c3975_1
- libev=4.33=h516909a_1
- libevent=2.1.10=hcdb4288_2
- libfaiss=1.6.3=h328c4c8_1_cuda
- libffi=3.3=he6710b0_2
- libgcc-ng=9.3.0=h24d8f2e_16
- libgfortran-ng=7.3.0=hdf63c60_0
- libgomp=9.3.0=h24d8f2e_16
- libhwloc=2.1.0=h3c4fd83_0
- libiconv=1.16=h516909a_0
- liblapack=3.8.0=17_openblas
- libllvm10=10.0.1=hbcb73fb_5
- libnghttp2=1.41.0=h8cfc5f6_2
- libopenblas=0.3.10=h5a2b251_0
- libpng=1.6.37=hbc83047_0
- libprotobuf=3.12.4=hd408876_0
- librmm=0.15.0=cuda11.0_g8005ca5_0
- libssh2=1.9.0=h1ba5d50_1
- libstdcxx-ng=9.1.0=hdf63c60_0
- libthrift=0.13.0=hbe8ec66_6
- libtiff=4.1.0=h2733197_1
- libwebp-base=1.1.0=h516909a_3
- libxml2=2.9.10=h68273f3_2
- llvmlite=0.34.0=py38h269e1b5_4
- locket=0.2.0=py38_1
- lz4-c=1.9.2=he6710b0_1
- mako=1.1.3=pyh9f0ad1d_0
- markupsafe=1.1.1=py38h7b6447c_0
- msgpack-python=1.0.0=py38hfd86e86_1
- nccl=2.7.8.1=h4962215_0
- ncurses=6.2=he6710b0_1
- numba=0.51.2=py38h0573a6f_1
- numpy=1.19.1=py38hbc27379_2
- olefile=0.46=py_0
- openssl=1.1.1h=h7b6447c_0
- packaging=20.4=py_0
- pandas=1.1.1=py38he6710b0_0
- parquet-cpp=1.5.1=2
- partd=1.1.0=py_0
- pillow=7.2.0=py38hb39fc2d_0
- pip=20.2.2=py38_0
- protobuf=3.12.4=py38h950e882_0
- psutil=5.7.2=py38h7b6447c_0
- pyarrow=0.17.1=py38h1234567_11_cuda
- pycparser=2.20=pyh9f0ad1d_2
- pyopenssl=19.1.0=py_1
- pyparsing=2.4.7=py_0
- pysocks=1.7.1=py38h32f6830_1
- python=3.8.5=h7579374_1
- python-dateutil=2.8.1=py_0
- python-editor=1.0.4=py_0
- python_abi=3.8=1_cp38
- pytz=2020.1=py_0
- pyyaml=5.3.1=py38h7b6447c_1
- querystring_parser=1.2.4=py_0
- re2=2020.07.06=he1b5a44_1
- readline=8.0=h7b6447c_0
- requests=2.24.0=pyh9f0ad1d_0
- rmm=0.15.0=cuda_11.0_py38_g8005ca5_0
- scikit-learn=0.23.2=py38h0573a6f_0
- scipy=1.5.2=py38h8c5af15_0
- setuptools=49.6.0=py38_0
- simplejson=3.17.2=py38h1e0a361_0
- six=1.15.0=py_0
- smmap=3.0.4=pyh9f0ad1d_0
- snappy=1.1.8=he6710b0_0
- sortedcontainers=2.2.2=py_0
- spdlog=1.8.0=hfd86e86_1
- sqlite=3.33.0=h62c20be_0
- sqlparse=0.3.1=py_0
- tabulate=0.8.7=pyh9f0ad1d_0
- tbb=2020.3=hfd86e86_0
- tblib=1.7.0=py_0
- threadpoolctl=2.1.0=pyh5ca1d4c_0
- thrift-compiler=0.13.0=hbe8ec66_6
- thrift-cpp=0.13.0=6
- tk=8.6.10=hbc83047_0
- toolz=0.10.0=py_0
- tornado=6.0.4=py38h7b6447c_1
- treelite=0.92=py38h4e709cc_2
- typing_extensions=3.7.4.3=py_0
- ucx=1.8.1+g6b29558=cuda11.0_0
- ucx-py=0.15.0+g6b29558=py38_0
- urllib3=1.25.10=py_0
- websocket-client=0.57.0=py38h32f6830_2
- werkzeug=1.0.1=pyh9f0ad1d_0
- wheel=0.35.1=py_0
- xz=5.2.5=h7b6447c_0
- yaml=0.2.5=h7b6447c_0
- zict=2.0.0=py_0
- zlib=1.2.11=h7b6447c_3
- zstd=1.4.5=h9ceee32_0
- pip:
- alembic==1.4.1
- azure-core==1.8.1
- azure-storage-blob==12.5.0
- isodate==0.6.0
- mlflow==1.11.0
- msrest==0.6.19
- oauthlib==3.1.0
- prometheus-client==0.8.0
- prometheus-flask-exporter==0.18.0
- requests-oauthlib==1.3.0
- sqlalchemy==1.3.13
- treelite-runtime==0.92
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/local_environment | rapidsai_public_repos/cloud-ml-examples/mlflow/local_environment/notebooks/databricks_mlflow_rapids.ipynb | import time
import subprocess
import sys
import threading
from queue import Queue, Empty
from functools import partial
import mlflow
import mlflow.sklearn
from cuml.metrics.accuracy import accuracy_score
from cuml.preprocessing.model_selection import train_test_split
from cuml.ensemble import RandomForestClassifier
from hyperopt import fmin, tpe, hp, Trials, STATUS_OKimport os
USER_NAME = ""
ACCOUNT_ID = ""
ACCOUNT_TOKEN = ""
experiment = "rapids_mlflow"
dbvars = {
"MLFLOW_EXPERIMENT_NAME": f"/Users/{USER_NAME}/{experiment}",
"MLFLOW_TRACKING_URI": f"databricks",
"DATABRICKS_HOST": f"https://{ACCOUNT_ID}.cloud.databricks.com",
"DATABRICKS_TOKEN": f"{ACCOUNT_TOKEN}"
}
def set_databricks_env():
for k, v in dbvars.items():
os.environ[k] = v
mlflow.set_experiment(f"/Users/{USER_NAME}/{experiment}")
set_databricks_env()def load_data(fpath):
"""
Simple helper function for loading data to be used by CPU/GPU models.
:param fpath: Path to the data to be ingested
:return: DataFrame wrapping the data at [fpath]. Data will be in either a Pandas or RAPIDS (cuDF) DataFrame
"""
import cudf
df = cudf.read_parquet(fpath)
return dfdef _train(params, fpath, hyperopt=False):
"""
:param params: hyperparameters. Its structure is consistent with how search space is defined. See below.
:param fpath: Path or URL for the training data used with the model.
:param mode: Hardware backend to use for training [CPU|GPU]
:param hyperopt: Use hyperopt for hyperparameter search during training.
:return: dict with fields 'loss' (scalar loss) and 'status' (success/failure status of run)
"""
max_depth, max_features, n_estimators = params
max_depth, max_features, n_estimators = int(max_depth), float(max_features), int(n_estimators)
df = load_data(fpath)
X = df.drop(["ArrDelayBinary"], axis=1)
y = df["ArrDelayBinary"].astype('int32')
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
mod = RandomForestClassifier(max_depth=max_depth, max_features=max_features, n_estimators=n_estimators)
acc_scorer = accuracy_score
mod.fit(X_train, y_train)
preds = mod.predict(X_test)
acc = acc_scorer(y_test, preds)
mlparams = {"max_depth": str(max_depth),
"max_features": str(max_features),
"n_estimators": str(n_estimators),
}
mlflow.log_params(mlparams)
mlmetrics = {"accuracy": acc}
mlflow.log_metrics(mlmetrics)
if (not hyperopt):
return mod
return {'loss': acc, 'status': STATUS_OK}
def train(params, fpath, hyperopt=False):
"""
Proxy function used to call _train
:param params: hyperparameters. Its structure is consistent with how search space is defined. See below.
:param fpath: Path or URL for the training data used with the model.
:param hyperopt: Use hyperopt for hyperparameter search during training.
:return: dict with fields 'loss' (scalar loss) and 'status' (success/failure status of run)
"""
with mlflow.start_run(nested=True):
return _train(params, fpath, hyperopt)PATH_TO_CONDA_DATA = ""
PATH_TO_AIRLINE_DATA = "rapidsai-cloud-ml-sample-data.s3-us-west-2.amazonaws.com"
algorithm = 'tpe'
conda_env = f'https://{PATH_TO_CONDA_DATA}/conda.yaml'
fpath = f'https://{PATH_TO_AIRLINE_DATA}/airline_small.parquet'
search_space = [
hp.uniform('max_depth', 5, 20),
hp.uniform('max_features', 0., 1.0),
hp.uniform('n_estimators', 150, 1000)
]
trials = Trials()
algorithm = tpe.suggest if algorithm == 'tpe' else None
fn = partial(train, fpath=fpath, hyperopt=True)
experid = 0
with mlflow.start_run():
mlflow.set_tag("mlflow.runName", "RAPIDS-Hyperopt-Databricks")
argmin = fmin(fn=fn,
space=search_space,
algo=algorithm,
max_evals=2,
trials=trials)
print("===========")
fn = partial(train, fpath=fpath, hyperopt=False)
final_model = fn(tuple(argmin.values()))
conda_data = ""
if (conda_env.startswith("http")):
import requests
resp = requests.get(conda_env)
conda_data = str(resp.text)
else:
with open(conda_env, 'r') as reader:
conda_data = reader.read()
with open("conda.yaml", 'w') as writer:
writer.write(conda_data)
mlflow.sklearn.log_model(final_model,
artifact_path="rapids_mlflow_test",
registered_model_name="rapids_mlflow_test",
conda_env='conda.yaml')
client = mlflow.tracking.MlflowClient()
latest_model = dict(client.search_model_versions("name='rapids_mlflow_test'")[0])
latest_model_source = latest_model['source']
retries = 0
while(True):
if (retries > 1):
raise RuntimeError("Failed to update registered model status.")
try:
# We need to wait for the model to be registered
time.sleep(10)
client.transition_model_version_stage(
name="rapids_mlflow_test",
version=latest_model['version'],
stage="Production")
print(f"Successfully registered model version {latest_model['version']}, as production.")
break
except Exception as e:
print(e, flush=True)
retries += 1def queue_descriptor_output(out, queue):
for line in iter(out.readline, b''):
queue.put(line)
out.close()
def follow_subprocess(cmd, timeout=1000, line_timeout=60.00):
p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
q = Queue()
t = threading.Thread(target=queue_descriptor_output, args=(p.stdout, q))
t.daemon = True
t.start()
elapsed = 0
line_elapsed = 0
last_line_time = time.perf_counter()
while (p.poll() is None and elapsed < timeout and line_elapsed < line_timeout):
try:
time.sleep(2)
elapsed += 2
while (True):
line = q.get(timeout=0.1)
line_elapsed = 0
last_line_time = time.perf_counter()
sys.stdout.write(line.decode())
except Empty:
line_elapsed = (time.perf_counter() - last_line_time)
except KeyboardInterrupt:
sys.stderr.write("\nCaught ctrl+c, killing subprocess ({})\n".format(' '.join(cmd)))
p.kill()
raise
try:
p.kill()
except:
pass
t.join(2)
## Drain any remaining text
try:
while (True):
line = q.get(timeout=0.1)
sys.stdout.write(line)
except Empty:
passport = 55755
host = 'localhost'
command = f"mlflow models serve -m {latest_model_source} -p {port} -h {host}".split()
kwargs = { "cmd": command, "timeout":float('Inf'), "line_timeout": float('Inf') }
threading.Thread(target=follow_subprocess, kwargs=kwargs).start()
time.sleep(30)import json
import requests
headers = {
"Content-Type": "application/json",
"format": "pandas-split"
}
data = {
"columns": ["Year", "Month", "DayofMonth", "DayofWeek", "CRSDepTime", "CRSArrTime", "UniqueCarrier", "FlightNum", "ActualElapsedTime", "Origin", "Dest", "Distance", "Diverted"],
"data": [[1987, 10, 1, 4, 1, 556, 0, 190, 247, 202, 162, 1846, 0]]
}
while (True):
try:
resp = requests.post(url=f"http://{host}:{port}/invocations", data=json.dumps(data), headers=headers)
print(f'Classification: {"ON-Time" if resp.text == "[0.0]" else "LATE"}')
break
except Exception as e:
errmsg = f"Caught exception attempting to call model endpoint: {e}"
print(f"{errmsg}", end='')
print(f"Sleeping")
time.sleep(20) | 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/local_environment | rapidsai_public_repos/cloud-ml-examples/mlflow/local_environment/src/sample_server_query.sh | curl -X POST -H "Content-Type:application/json; format=pandas-split" --data '{"columns":["Year", "Month", "DayofMonth", "DayofWeek", "CRSDepTime", "CRSArrTime", "UniqueCarrier", "FlightNum", "ActualElapsedTime", "Origin" , "Dest", "Distance", "Diverted"],"data":[[1987, 10, 1, 4, 1, 556, 0, 190, 247, 202, 162, 1846, 0]]}' http://127.0.0.1:55756/invocations
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/local_environment/src | rapidsai_public_repos/cloud-ml-examples/mlflow/local_environment/src/rf_test/train_simple.py | """Simple example integrating cuML with MLFlow"""
import argparse
import mlflow
import mlflow.sklearn
from cuml.metrics.accuracy import accuracy_score
from cuml.model_selection import train_test_split
from cuml.ensemble import RandomForestClassifier
def load_data(fpath):
"""
Simple helper function for loading data to be used by CPU/GPU models.
:param fpath: Path to the data to be ingested
:return: DataFrame wrapping the data at [fpath]. Data will be in either a Pandas or RAPIDS (cuDF) DataFrame
"""
import cudf
df = cudf.read_parquet(fpath)
X = df.drop(["ArrDelayBinary"], axis=1)
y = df["ArrDelayBinary"].astype("int32")
return train_test_split(X, y, test_size=0.2)
def train(fpath, max_depth, max_features, n_estimators):
"""
:param params: hyperparameters. Its structure is consistent with how search space is defined. See below.
:param fpath: Path or URL for the training data used with the model.
:param max_depth: RF max_depth parameter
:param max_features: RF max_features parameter
:param n_estimators: RF n_estimators parameter
:return: trained model
"""
X_train, X_test, y_train, y_test = load_data(fpath)
mod = RandomForestClassifier(
max_depth=max_depth, max_features=max_features, n_estimators=n_estimators
)
mod.fit(X_train, y_train)
preds = mod.predict(X_test)
acc = accuracy_score(y_test, preds)
mlparams = {
"max_depth": str(max_depth),
"max_features": str(max_features),
"n_estimators": str(n_estimators),
}
mlflow.log_params(mlparams)
mlflow.log_metric("accuracy", acc)
mlflow.sklearn.log_model(mod, "saved_models")
return mod
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument("--algo", default="tpe", choices=["tpe"], type=str)
parser.add_argument("--conda_env", required=True, type=str)
parser.add_argument("--fpath", required=True, type=str)
parser.add_argument("--n_estimators", type=int, default=100)
parser.add_argument("--max_features", type=float, default=1.0)
parser.add_argument("--max_depth", type=int, default=12)
args = parser.parse_args()
experid = 0
artifact_path = "Airline-Demo"
artifact_uri = None
experiment_name = "RAPIDS-MLflow"
experiment_id = None
with mlflow.start_run(run_name="RAPIDS-MLflow"):
model = train(args.fpath, args.max_depth, args.max_features, args.n_estimators)
mlflow.sklearn.log_model(
model,
artifact_path=artifact_path,
registered_model_name="rapids_mlflow_cli",
conda_env=args.conda_env,
)
artifact_uri = mlflow.get_artifact_uri(artifact_path=artifact_path)
print("Model uri: %s" % artifact_uri)
| 0 |
rapidsai_public_repos/cloud-ml-examples/mlflow/local_environment/src | rapidsai_public_repos/cloud-ml-examples/mlflow/local_environment/src/rf_test/train.py | """Hyperparameter optimization with cuML, hyperopt, and MLFlow"""
import argparse
from functools import partial
import mlflow
import mlflow.sklearn
from cuml.metrics.accuracy import accuracy_score
from cuml.model_selection import train_test_split
from cuml.ensemble import RandomForestClassifier
from hyperopt import fmin, tpe, hp, Trials, STATUS_OK
def load_data(fpath):
"""
Simple helper function for loading data to be used by CPU/GPU models.
:param fpath: Path to the data to be ingested
:return: DataFrame wrapping the data at [fpath]. Data will be in either a Pandas or RAPIDS (cuDF) DataFrame
"""
import cudf
df = cudf.read_parquet(fpath)
X = df.drop(["ArrDelayBinary"], axis=1)
y = df["ArrDelayBinary"].astype("int32")
return train_test_split(X, y, test_size=0.2)
def _train(params, fpath, hyperopt=False):
"""
:param params: hyperparameters. Its structure is consistent with how search space is defined. See below.
:param fpath: Path or URL for the training data used with the model.
:param hyperopt: Use hyperopt for hyperparameter search during training.
:return: dict with fields 'loss' (scalar loss) and 'status' (success/failure status of run)
"""
max_depth, max_features, n_estimators = params
max_depth, max_features, n_estimators = (int(max_depth), float(max_features), int(n_estimators))
X_train, X_test, y_train, y_test = load_data(fpath)
mod = RandomForestClassifier(
max_depth=max_depth, max_features=max_features, n_estimators=n_estimators
)
mod.fit(X_train, y_train)
preds = mod.predict(X_test)
acc = accuracy_score(y_test, preds)
mlparams = {
"max_depth": str(max_depth),
"max_features": str(max_features),
"n_estimators": str(n_estimators),
}
mlflow.log_params(mlparams)
mlflow.log_metric("accuracy", acc)
mlflow.sklearn.log_model(mod, "saved_models")
if not hyperopt:
return mod
return {"loss": acc, "status": STATUS_OK}
def train(params, fpath, hyperopt=False):
"""
Proxy function used to call _train
:param params: hyperparameters. Its structure is consistent with how search space is defined. See below.
:param fpath: Path or URL for the training data used with the model.
:param hyperopt: Use hyperopt for hyperparameter search during training.
:return: dict with fields 'loss' (scalar loss) and 'status' (success/failure status of run)
"""
with mlflow.start_run(nested=True):
return _train(params, fpath, hyperopt)
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument("--algo", default="tpe", choices=["tpe"], type=str)
parser.add_argument("--conda-env", required=True, type=str)
parser.add_argument("--fpath", required=True, type=str)
args = parser.parse_args()
search_space = [
hp.uniform("max_depth", 5, 20),
hp.uniform("max_features", 0.1, 1.0),
hp.uniform("n_estimators", 150, 1000),
]
trials = Trials()
algorithm = tpe.suggest if args.algo == "tpe" else None
fn = partial(train, fpath=args.fpath, hyperopt=True)
experid = 0
artifact_path = "Airline-Demo"
artifact_uri = None
with mlflow.start_run(run_name="RAPIDS-Hyperopt"):
argmin = fmin(fn=fn, space=search_space, algo=algorithm, max_evals=2, trials=trials)
print("===========")
fn = partial(train, fpath=args.fpath, hyperopt=False)
final_model = fn(tuple(argmin.values()))
mlflow.sklearn.log_model(
final_model,
artifact_path=artifact_path,
registered_model_name="rapids_mlflow_cli",
conda_env="envs/conda.yaml",
)
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/README.md | ## Inferencing with Triton Inferencing Server with RAPIDS cuML FIL backend on Kubernetes
This folder contains examples to run the Triton Inferencing Server on a Kubernetes Server with a custom RAPIDS cuML FIL backend.
1. The directory [GCP](./GCP) contains examples on Google Kubernetes Engine.
2. The directory [AWS](./AWS) contains examples on Amazon Elastic Kubernetes Services.
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/README.md | ## Deploying an autoscaling Triton service utilizing a custom plugin for the RAPIDS cuML Forest Inference Library (FIL).
#### Adapted from Dong Meng's `gke-marketplace-app` in the [triton-inference-server](https://github.com/triton-inference-server/server/tree/master/deploy/gke-marketplace-app) repository.
### Overview
This example will illustrate the workflow to deploy a horizontally scaled Triton service, with a non-trival custom
backend for accelerated forest model inference. It is assumed that you have an existing account with sufficient compute
and GPU quota, a correctly configured kubectl utility, and a [properly configured](https://github.com/rapidsai/cloud-ml-examples/blob/main/mlflow/docker_environment/DetailedConfig.md#create-a-storage-bucket-and-attach-your-service-account) GCP bucket.
For simplicity, this process will be demonstrated using the Google Kubernetes Engine (GKE) platform. But should be
straight-forward to adapt to any desired Kubernetes environment with Istio and the stackdriver custom metrics adapter.
When referring to configuration parameters or elements that are `specific to your environment`, they will be represented
as linux-style environment variables ex. ${YOUR_CONFIG_PARAM}.
Specific parameters used here include:
- YOUR_PROJECT_ID : Your GCP [Project ID](https://support.google.com/googleapi/answer/7014113?hl=en).
- YOUR_GCR_PATH : We assume a GCR model location of `gcr.io/${YOUR_PROJECT_ID}/${YOUR_GCR_PATH}/<..models..>`.
- YOUR_BUCKET_PATH : Path to your personal, writable, GCP bucket.
- YOUR_CLUSTER_ID : Name of your GKE cluster.
- YOUR_CLUSTER_ZONE : Zone where the cluster will be deployed.
#### Pre-requisites
- GCP Account with the following permissions
- container.clusterRoleBindings.create
- container.clusterRoles.update
- container.roleBindings.create
- System Software
- `docker`
- `gcloud` `gsutil`
- `helm` `kubectl`
- Python
```shell
pip install nvidia-pyindex
pip install tritonclient[all]
```
### Obtain the Triton FIL plugin, build the triton host container, and push to GCR
Note: as of this writing, the [FIL backend plugin](https://github.com/wphicks/triton_fil_backend) is considered
experimental / preview-quality.
```shell
git clone git@github.com:wphicks/triton_fil_backend.git
cd triton_fil_backend
docker build --tag gcr.io/${YOUR_PROJECT_ID}/${YOUR_GCR_PATH}/triton_fil --filename ops/Dockerfile .
docker push gcr.io/${YOUR_PROJECT_ID}/${YOUR_GCR_PATH}/triton_fil:latest
```
### Create a Triton model registry entry, or use the provided example
A sample XGBoost model, along with its .pbtext defintion is provided in ./model_repository. The layout structure and
requirements are defined in the [Triton server docs](https://github.com/triton-inference-server/server/blob/master/docs/model_configuration.md),
and a brief introduction is also provided with the [FIL backend implementation](https://github.com/wphicks/triton_fil_backend#triton-inference-server-fil-backend).
This is what will be referenced for the purpose of this demo; however, feel free to add additional models, following the
same structure, and they will be included in subsequent steps.
```shell
gsutil cp -r ./model_repository gs://${YOUR_BUCKET_PATH}/triton/
gsutil ls gs://${YOUR_BUCKET_PATH}/triton/
gs://${YOUR_BUCKET_PATH}/triton/
gs://${YOUR_BUCKET_PATH}/triton/model_repository/
```
### Configure GKE cluster
This step is equivalent to the triton-inference-server sample, we just need to create a cluster that will host our
Triton service, and has a GPU node pool with enough nodes to illustrate horizontal scaling.
```shell
gcloud beta container clusters create ${YOUR_CLUSTER_ID} \
--addons=HorizontalPodAutoscaling,HttpLoadBalancing,Istio \
--machine-type=n1-standard-8 \
--node-locations=${YOUR_CLUSTER_ZONE} \
--zone=${YOUR_CLUSTER_ZONE} \
--subnetwork=default \
--scopes=cloud-platform \
--num-nodes=1
# add GPU node pools, user can modify number of node based on workloads
gcloud container node-pools create gpu-pool \
--project ${YOUR_PROJECT_ID} \
--zone ${YOUR_CLUSTER_ZONE} \
--cluster ${YOUR_CLUSTER_ID} \
--num-nodes 2 \
--accelerator type=nvidia-tesla-t4,count=1 \
--enable-autoscaling --min-nodes 2 --max-nodes 3 \
--machine-type n1-standard-4 \
--disk-size=100 \
--scopes cloud-platform \
--verbosity error
# so that you can run kubectl locally to the cluster
gcloud container clusters get-credentials ${YOUR_CLUSTER_ID} --project ${YOUR_PROJECT_ID} --zone ${YOUR_CLUSTER_ZONE}
# create a shared secret for the gcp bucket where your custom model lives
kubectl create secret generic gcsfs-creds --from-file=./keyfile.json
# deploy NVIDIA device plugin for GKE to prepare GPU nodes for driver install
kubectl apply -f https://raw.githubusercontent.com/GoogleCloudPlatform/container-engine-accelerators/master/nvidia-driver-installer/cos/daemonset-preloaded.yaml
# make sure you can run kubectl locally to access the cluster
kubectl create clusterrolebinding cluster-admin-binding --clusterrole cluster-admin --user "$(gcloud config get-value account)"
# enable stackdriver custom metrics adaptor
kubectl apply -f https://raw.githubusercontent.com/GoogleCloudPlatform/k8s-stackdriver/master/custom-metrics-stackdriver-adapter/deploy/production/adapter.yaml
```
### Install Triton service using helm.
```shell
helm install triton ./helm/chart/triton \
--set modelRepositoryPath="gs://${YOUR_BUCKET_PATH}/triton/model_repository" \
--set image.repository="${YOUR_PROJECT_ID}/${YOUR_GCR_PATH}/triton_fil"
```
When finished, you can delete the triton deployment using
```shell
helm uninstall triton
```
### Exploring inference
At this point, your triton inference cluster should be up and running or in process of coming up. Now we can submit some
test data to our running server. The process for doing this, assuming the default model, is illustrated in the jupyter
notebook `interact_with_triton.ipynb`. | 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/interact_with_triton.ipynb | import os
import numpy
import subprocess
import sys
import time
import tritonclient.http as triton_http
import tritonclient.grpc as triton_grpchttp_port_cmd = "kubectl -n istio-system get service istio-ingressgateway -o jsonpath='{.spec.ports[?(@.name==\"http2\")].port}'"
grpc_port_cmd = "kubectl -n istio-system get service istio-ingressgateway -o jsonpath='{.spec.ports[?(@.name==\"tcp\")].port}'"
host_cmd = "kubectl -n istio-system get service istio-ingressgateway -o jsonpath='{.status.loadBalancer.ingress[0].ip}'"
http_port = subprocess.check_output(http_port_cmd.split()).decode('utf-8').replace("'", "")
grpc_port = subprocess.check_output(grpc_port_cmd.split()).decode('utf-8').replace("'", "")
host = subprocess.check_output(host_cmd.split()).decode('utf-8').replace("'", "")
print(host, http_port)
print(host, grpc_port)# Set up both HTTP and GRPC clients. Note that the GRPC client is generally
# somewhat faster.
# Generate dummy data to classify
features = 500
samples = 10_000
data = numpy.random.rand(samples, features).astype('float32')http_client = triton_http.InferenceServerClient(
url=f'{host}:{http_port}',
verbose=False,
concurrency=12
)
while (not (http_client.is_server_ready() or http_client.is_model_ready('fil'))):
print("Waiting on server ready")
time.sleep(5)
print(f"Is Server Ready: {http_client.is_server_ready()}")
print(f"Is FIL model ready: {http_client.is_model_ready('fil')}")# Set up Triton input and output objects for both HTTP and GRPC
triton_input_http = triton_http.InferInput(
'input__0',
(samples, features),
'FP32'
)
triton_input_http.set_data_from_numpy(data, binary_data=True)
triton_output_http = triton_http.InferRequestedOutput(
'output__0',
binary_data=True
)
# Submit inference requests (both HTTP and GRPC)
request_http = http_client.infer(
'fil',
model_version='1',
inputs=[triton_input_http],
outputs=[triton_output_http]
)result_http = request_http.as_numpy('output__0')
result_httpgrpc_client = triton_grpc.InferenceServerClient(
url=f'{host}:{grpc_port}',
verbose = False
)
while (not (grpc_client.is_server_ready() or grpc_client.is_model_ready('fil'))):
print("Waiting on server ready")
time.sleep(5)
print(f"Is Server Ready: {grpc_client.is_server_ready()}")
print(f"Is FIL model ready: {grpc_client.is_model_ready('fil')}")triton_input_grpc = triton_grpc.InferInput(
'input__0',
[samples, features],
'FP32'
)
triton_input_grpc.set_data_from_numpy(data)
triton_output_grpc = triton_grpc.InferRequestedOutput('output__0')
request_grpc = grpc_client.infer(
'fil',
model_version='1',
inputs=[triton_input_grpc],
outputs=[triton_output_grpc]
)result_grpc = request_grpc.as_numpy('output__0')
result_grpc# Check that we got the same result with both GRPC and HTTP
numpy.testing.assert_almost_equal(result_http, result_grpc) | 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/model_repository | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/model_repository/fil/config.pbtxt | name: "fil"
backend: "fil"
max_batch_size: 1048576
input [
{
name: "input__0"
data_type: TYPE_FP32
dims: [ 500 ]
}
]
output [
{
name: "output__0"
data_type: TYPE_FP32
dims: [ 2 ]
}
]
instance_group [{ kind: KIND_GPU }]
dynamic_batching {
preferred_batch_size: [1, 2, 4, 8, 16, 32, 64, 128, 1024, 131072, 1048576]
max_queue_delay_microseconds: 30000
}
parameters [
{
key: "algo"
value: { string_value: "ALGO_AUTO" }
},
{
key: "storage_type"
value: { string_value: "AUTO" }
},
{
key: "output_class"
value: { string_value: "true" }
},
{
key: "threshold"
value: { string_value: "0.5" }
},
{
key: "blocks_per_sm"
value: { string_value: "0" }
},
{
key: "predict_proba"
value: { string_value: "true" }
},
{
key: "model_type"
value: { string_value: "xgboost_json" }
}
] | 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/model_repository/fil | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/model_repository/fil/1/xgboost.json | {"learner":{"attributes":{"scikit_learn":"{\"use_label_encoder\": false, \"n_estimators\": 500, \"objective\": \"binary:logistic\", \"max_depth\": 7, \"learning_rate\": null, \"verbosity\": null, \"booster\": null, \"tree_method\": \"gpu_hist\", \"gamma\": null, \"min_child_weight\": null, \"max_delta_step\": null, \"subsample\": null, \"colsample_bytree\": null, \"colsample_bylevel\": null, \"colsample_bynode\": null, \"reg_alpha\": null, \"reg_lambda\": null, \"scale_pos_weight\": null, \"base_score\": null, \"missing\": NaN, \"num_parallel_tree\": null, \"kwargs\": {\"eval_metric\": \"error\", \"predictor\": \"gpu_predictor\"}, \"random_state\": null, \"n_jobs\": null, \"monotone_constraints\": null, \"interaction_constraints\": null, \"importance_type\": \"gain\", \"gpu_id\": null, \"validate_parameters\": null, \"classes_\": [0, 1], \"n_classes_\": 2, \"_features_count\": 500, \"n_features_in_\": 500, \"type\": \"XGBClassifier\"}"},"gradient_booster":{"model":{"gbtree_model_param":{"num_trees":"500","size_leaf_vector":"0"},"tree_info":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"trees":[{"base_weights":[-4.3824703E-2,2.3417722E-1,-5.106383E-1,-1.8068536E-1,6.539683E-1,6.486486E-1,-7.8431374E-1,-4.651163E-1,9.2537314E-1,1.3043479E0,-3.75E-1,-8.5714287E-1,7.9037803E-1,-1.0229008E0,5.833333E-1,-7.348837E-1,8.0851066E-1,-1.20000005E-1,1.6E0,-1.4054054E0,-5.084746E-1,1.0989011E-1,1.0784314E0,1.0243902E0,-3.818182E-1,-4.8000002E-1,6.666667E-1,-3.2307696E-1,1.3103448E0,-1.04E0,3.818182E-1,-9.183673E-1,9.5652175E-1,5.3333336E-1,-0E0,-7.5E-1,1.2083334E0,-4.8000002E-1,2.0000002E-1,1.6226416E0,-3.3333337E-1,3E-1,-0E0,6.885246E-1,-9.411765E-1,1.5E0,-4.6153846E-1,-9.4505495E-1,8.3870965E-1,-4E-1,1.483871E0,-1.5820895E0,2.2222222E-1,1.3333334E-1,-1.2E0,1.5E-1,-4.0000004E-1,-5E-1,5.714286E-1,-1.6615385E0,1.5E-1,4.0000004E-1,-3.6E-1,-0E0,5.1111114E-1,3.3333337E-1,-3.6E-1,-3.3333337E-1,1.3846154E0,7.692308E-1,-1.1707317E0,-3.3333337E-1,1.5E-1,1.2727273E0,-6.666667E-1,-0E0,5.1428574E-1,5.5E-1,-6.666667E-2,4.7368425E-1,-3E-1,1.3757225E0,-2.6086956E-1,4.4000003E-1,-1.4117648E-1,3.6E-1,-1.5E-1,-4.0000004E-1,3E-1,3.2307696E-1,-4.2857146E-1,4.666667E-1,-0E0,-3.8823533E-1,3E-1,-0E0,3.3333337E-1,5E-1,-1.0000001E-1,4.909091E-1,-0E0,-4.4615388E-1,-8.5714296E-2,-5.2380955E-1,1.5E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,false,true,true,true,true,true,true,false,false,true,false,true,true,false,true,true,false,false,true,true,false,false,true,false,false,false,true,true,true,true,true,true,true,true,true,true,true,true,false,false,false,true,true,false,false,false,false,false,false,false,false,true,true,true,false,false,true,true,false,false,false,false,false,false,true,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":0,"left_children":[1,3,5,7,11,9,13,15,27,17,29,25,21,19,23,31,49,-1,33,51,47,43,35,45,-1,-1,41,-1,39,37,-1,53,79,-1,-1,55,81,-1,-1,77,-1,-1,-1,73,57,75,71,69,67,65,63,59,61,83,101,-1,-1,-1,85,103,-1,-1,-1,-1,-1,-1,-1,-1,99,95,87,-1,-1,97,93,-1,-1,-1,-1,-1,-1,91,89,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[3.2693123E1,2.7634745E1,3.0331966E1,2.5676834E1,1.6910229E1,1.2906434E1,2.555886E1,2.2755268E1,1.4322499E1,6.434782E0,1.0089546E1,8.990477E0,1.4141716E1,1.2173851E1,1.112731E1,1.7559492E1,1.0183664E1,0E0,2.8444443E0,1.099189E1,1.8143175E1,1.4484196E1,1.3019611E1,7.9362106E0,0E0,0E0,6.666666E-1,0E0,1.2768019E1,7.04E0,0E0,1.467347E1,8.581236E0,0E0,0E0,3.5833335E0,1.216346E1,0E0,0E0,5.5576515E0,0E0,0E0,0E0,1.2922008E1,1.0280111E1,2.5714283E0,2.58547E0,9.701956E0,9.787703E0,7.377778E0,2.5280762E0,6.3723297E0,8.711111E0,9.74902E0,9.749447E0,0E0,0E0,0E0,2.9571428E0,6.80867E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,2.2657347E0,8.547008E-1,8.569107E0,0E0,0E0,7.515152E0,6.784314E0,0E0,0E0,0E0,0E0,0E0,0E0,1.0700066E1,1.0520784E1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,13,13,12,12,14,14,11,11,8,8,10,10,15,15,18,18,22,22,29,29,28,28,26,26,21,21,23,23,20,20,16,16,19,19,31,31,35,35,44,44,51,51,52,52,50,50,49,49,48,48,47,47,46,46,43,43,45,45,39,39,32,32,36,36,53,53,58,58,70,70,82,82,81,81,74,74,69,69,73,73,68,68,54,54,59,59],"right_children":[2,4,6,8,12,10,14,16,28,18,30,26,22,20,24,32,50,-1,34,52,48,44,36,46,-1,-1,42,-1,40,38,-1,54,80,-1,-1,56,82,-1,-1,78,-1,-1,-1,74,58,76,72,70,68,66,64,60,62,84,102,-1,-1,-1,86,104,-1,-1,-1,-1,-1,-1,-1,-1,100,96,88,-1,-1,98,94,-1,-1,-1,-1,-1,-1,92,90,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[2.2912774E0,-1.9463277E-1,-8.009977E0,7.6190567E0,-1.6127809E0,3.0724077E0,7.940915E0,6.779244E0,-1.0607927E0,-7.5571384E0,7.1968293E-1,2.064278E0,-4.911548E0,-1.0237694E-2,7.989456E0,9.436388E0,-3.4294957E-1,-1.20000005E-1,1.5320612E0,1.3254604E0,3.2437477E0,3.1923423E0,-1.2095527E1,9.89423E-1,-3.818182E-1,-4.8000002E-1,-5.158195E-1,-3.2307696E-1,6.4011908E0,6.952167E-1,3.818182E-1,-6.7668897E-1,2.7658224E-1,5.3333336E-1,-0E0,-7.1275795E-1,1.2080219E0,-4.8000002E-1,2.0000002E-1,9.845266E0,-3.3333337E-1,3E-1,-0E0,4.1400275E0,7.9102165E-1,-9.11137E-1,-2.1021953E0,-1.1829063E0,-7.776553E0,-8.741593E-1,-1.1409214E1,1.1247731E1,1.0970592E-1,-7.7351093E-1,2.725833E0,1.5E-1,-4.0000004E-1,-5E-1,-9.1322947E-1,1.6601745E0,1.5E-1,4.0000004E-1,-3.6E-1,-0E0,5.1111114E-1,3.3333337E-1,-3.6E-1,-3.3333337E-1,3.8868237E-1,3.6968708E-1,1.142351E1,-3.3333337E-1,1.5E-1,7.551317E-1,6.917934E-1,-0E0,5.1428574E-1,5.5E-1,-6.666667E-2,4.7368425E-1,-3E-1,1.110732E0,-2.4737551E0,4.4000003E-1,-1.4117648E-1,3.6E-1,-1.5E-1,-4.0000004E-1,3E-1,3.2307696E-1,-4.2857146E-1,4.666667E-1,-0E0,-3.8823533E-1,3E-1,-0E0,3.3333337E-1,5E-1,-1.0000001E-1,4.909091E-1,-0E0,-4.4615388E-1,-8.5714296E-2,-5.2380955E-1,1.5E-1],"split_indices":[81,73,61,161,308,25,5,55,319,110,392,155,48,45,134,45,213,0,263,293,71,50,117,331,0,0,258,0,156,321,0,280,136,0,0,296,210,0,0,150,0,0,0,21,458,313,10,314,55,497,20,78,351,464,56,0,0,0,1,377,0,0,0,0,0,0,0,0,270,1,153,0,0,209,277,0,0,0,0,0,0,463,95,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[2.5E2,1.57E2,9.3E1,7.925E1,7.775E1,1.75E1,7.55E1,6.35E1,1.575E1,1.05E1,7E0,6E0,7.175E1,6.45E1,1.1E1,5.275E1,1.075E1,1.5E0,9E0,3.6E1,2.85E1,2.175E1,5E1,9.25E0,1.75E0,4E0,2E0,2.25E0,1.35E1,5.25E0,1.75E0,4.8E1,4.75E0,8E0,1E0,3E0,4.7E1,4E0,1.25E0,1.225E1,1.25E0,1E0,1E0,1.425E1,7.5E0,7E0,2.25E0,2.175E1,6.75E0,4E0,6.75E0,3.25E1,3.5E0,1.025E1,3.775E1,1E0,2E0,5E0,2.5E0,3.15E1,1E0,2E0,1.5E0,1E0,5.75E0,1.25E0,2.75E0,1.25E0,5.5E0,2.25E0,1.95E1,1.25E0,1E0,1E1,4.25E0,1E0,6E0,1.1E1,1.25E0,3.75E0,1E0,4.225E1,4.75E0,2.75E0,7.5E0,1.5E0,1E0,1.85E1,1E0,2.25E0,2.5E0,3.725E1,5E0,3.25E0,1E0,1E0,1.25E0,8E0,2E0,4.5E0,1E0,2.825E1,9.5E0,3.05E1,1E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"105","size_leaf_vector":"0"}},{"base_weights":[-2.8290551E-2,2.6683834E-1,-3.4635958E-1,4.5143503E-1,-4.173971E-1,-6.65618E-1,1.6550638E-1,9.0648246E-1,1.026992E-1,5.3418964E-2,-9.4665694E-1,-1.3146446E0,-1.1466595E-2,-2.2535798E-1,8.772819E-1,1.0390147E0,-4.1537002E-1,-6.2030953E-1,7.8025794E-1,-6.5627265E-1,6.045097E-1,6.480958E-1,-8.426851E-1,1.2710571E0,-6.146229E-1,-4.5015186E-1,-2.8978245E-2,2.7184364E-1,-3.928355E-1,5.845759E-1,-1.1237934E0,-2.8924716E-1,1.1806383E0,6.4334875E-1,-1.112724E0,-5.364387E-1,1.1981167E0,-1.2068408E0,3.262564E-1,3.7590718E-1,-2.123947E-1,-9.719347E-1,4.420892E-1,-3.2196996E-1,8.413729E-2,-6.133764E-2,4.511183E-1,-7.695366E-1,1.0868114E0,-4.209642E-1,2.7966425E-1,2.9289204E-1,-9.023835E-1,1.5846328E-1,-3.4088516E-1,-3.0252555E-1,4.135662E-1,-3.4767076E-1,2.2933856E-1,3.285761E-1,-6.201004E-2,-1.3588214E0,-1.2931542E-1,-3.43369E-1,2.2225711E-1,3.532104E-1,-1.1370815E0,3.1045794E-1,-2.2840035E-1,4.2795524E-1,-1.6069408E-1,-3.037842E-1,-5.2407514E-2,-3.172128E-1,7.7022177E-1,4.35956E-4,4.680775E-1,-4.1511643E-1,8.377433E-2,-3.4155136E-1,1.7129694E-1,1.3329766E0,-2.2223455E-1,-4.368762E-1,1.0764899E-1,-4.3385708E-1,1.1766021E-1,2.9820228E-1,-1.5796998E-1,3.3302438E-2,4.5110932E-1,-1.9752651E-1,3.070863E-1,2.2162558E-1,-3.2596385E-1,2.3762187E-1,-2.793841E-1,3.1829318E-1,-2.451822E-1,-4.0648365E-1,8.33005E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,true,true,false,true,true,true,false,true,true,true,true,true,true,false,true,true,true,true,true,false,true,false,false,false,false,true,true,true,true,true,true,true,true,true,false,false,false,true,true,false,false,false,false,false,true,false,true,false,true,false,false,true,false,false,false,false,false,true,true,false,false,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":1,"left_children":[1,3,5,7,11,9,13,15,27,17,29,25,21,19,23,31,53,33,55,51,47,35,49,45,43,-1,-1,41,-1,39,37,57,81,59,83,79,75,61,-1,-1,-1,77,73,-1,-1,-1,-1,71,69,-1,67,-1,65,-1,-1,63,-1,-1,-1,-1,-1,85,97,-1,-1,87,99,-1,-1,-1,-1,-1,-1,95,91,-1,-1,-1,-1,-1,-1,89,93,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[2.2611902E1,1.597698E1,1.9049671E1,1.5605116E1,9.847251E0,1.4514555E1,1.2806066E1,7.9281654E0,1.4002177E1,1.0797323E1,1.462989E1,1.7625628E0,1.1320245E1,1.1142862E1,1.0339102E1,7.7113724E0,3.7387156E0,7.8634796E0,7.267145E0,9.021069E0,7.987795E0,8.15001E0,6.1358094E0,4.838415E0,1.9723202E0,0E0,0E0,1.1408422E1,0E0,6.070595E0,9.8549E0,5.0810595E0,7.904316E0,1.7071105E0,4.8863277E0,3.5467863E0,3.541852E0,7.243828E0,0E0,0E0,0E0,3.763702E0,1.1778338E1,0E0,0E0,0E0,0E0,3.1315613E-1,5.3950233E0,0E0,3.8339288E0,0E0,5.742777E0,0E0,0E0,4.874717E0,0E0,0E0,0E0,0E0,0E0,6.6255493E0,6.3581142E0,0E0,0E0,2.5627444E0,5.156006E0,0E0,0E0,0E0,0E0,0E0,0E0,1.0735413E1,1.2481497E1,0E0,0E0,0E0,0E0,0E0,0E0,6.5621147E0,4.4109697E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,13,13,12,12,14,14,11,11,8,8,10,10,15,15,17,17,21,21,30,30,29,29,27,27,24,24,23,23,20,20,22,22,19,19,16,16,18,18,31,31,33,33,37,37,55,55,52,52,50,50,48,48,47,47,42,42,36,36,41,41,35,35,32,32,34,34,61,61,65,65,81,81,74,74,82,82,73,73,62,62,66,66],"right_children":[2,4,6,8,12,10,14,16,28,18,30,26,22,20,24,32,54,34,56,52,48,36,50,46,44,-1,-1,42,-1,40,38,58,82,60,84,80,76,62,-1,-1,-1,78,74,-1,-1,-1,-1,72,70,-1,68,-1,66,-1,-1,64,-1,-1,-1,-1,-1,86,98,-1,-1,88,100,-1,-1,-1,-1,-1,-1,96,92,-1,-1,-1,-1,-1,-1,90,94,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[5.785241E-1,6.1989408E0,2.093761E0,-1.4496181E0,-6.1934257E0,-3.3271177E0,3.932939E0,1.1823459E0,1.0539654E1,-1.280942E-1,-6.943634E1,6.4332128E0,3.721261E-1,2.05485E0,5.0306325E0,-1.112236E1,2.0070326E0,-1.0305706E0,-2.6544113E0,-1.2972939E0,-5.98519E0,-5.700713E-1,1.4516401E0,-1.1579763E1,2.1083403E-1,-4.5015186E-1,-2.8978245E-2,-8.849676E0,-3.928355E-1,3.5978248E0,1.4644355E1,-3.4819694E0,8.895502E0,1.2632985E0,5.9471446E-1,3.568717E-1,-8.2452214E-1,6.7520375E0,3.262564E-1,3.7590718E-1,-2.123947E-1,1.468729E0,-2.9856353E0,-3.2196996E-1,8.413729E-2,-6.133764E-2,4.511183E-1,-1.3630092E0,9.500473E0,-4.209642E-1,3.3921003E-1,2.9289204E-1,-6.3520336E0,1.5846328E-1,-3.4088516E-1,-5.598025E-1,4.135662E-1,-3.4767076E-1,2.2933856E-1,3.285761E-1,-6.201004E-2,1.814569E0,-3.298215E0,-3.43369E-1,2.2225711E-1,-1.0422344E-1,8.329554E0,3.1045794E-1,-2.2840035E-1,4.2795524E-1,-1.6069408E-1,-3.037842E-1,-5.2407514E-2,-5.826699E0,-8.4342575E0,4.35956E-4,4.680775E-1,-4.1511643E-1,8.377433E-2,-3.4155136E-1,1.7129694E-1,-1.2725804E0,1.6506664E-1,-4.368762E-1,1.0764899E-1,-4.3385708E-1,1.1766021E-1,2.9820228E-1,-1.5796998E-1,3.3302438E-2,4.5110932E-1,-1.9752651E-1,3.070863E-1,2.2162558E-1,-3.2596385E-1,2.3762187E-1,-2.793841E-1,3.1829318E-1,-2.451822E-1,-4.0648365E-1,8.33005E-2],"split_indices":[159,164,47,86,108,69,16,259,110,103,166,11,65,5,8,36,158,296,158,255,33,389,73,20,5,0,0,153,0,53,103,55,142,55,427,407,323,73,0,0,0,137,41,0,0,0,0,154,15,0,181,0,126,0,0,260,0,0,0,0,0,328,81,0,0,389,138,0,0,0,0,0,0,105,56,0,0,0,0,0,0,408,186,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[2.3889725E2,1.2400082E2,1.1489643E2,9.792289E1,2.6077927E1,7.057646E1,4.431997E1,4.17917E1,5.613119E1,2.006083E1,5.0515633E1,7.4257793E0,1.8652147E1,2.905054E1,1.5269425E1,3.8177864E1,3.6138346E0,1.0486836E1,9.573993E0,1.9259232E1,9.79131E0,1.0522762E1,8.129385E0,1.2197922E1,3.0715034E0,6.2415004E0,1.1842792E0,5.0848198E1,5.282994E0,4.9993386E0,4.5516293E1,3.5780106E0,3.4599854E1,2.8546286E0,7.632208E0,3.3445578E0,7.1782045E0,4.4306572E1,1.2097201E0,3.3153749E0,1.6839639E0,5.5550632E0,4.5293137E1,1.8993189E0,1.1721845E0,1.6779829E0,1.0519938E1,2.3891697E0,7.4021406E0,5.2547746E0,2.8746107E0,2.1343107E0,1.712492E1,1.6809703E0,1.9328643E0,3.583324E0,5.990669E0,1.9055091E0,1.6725014E0,1.6754583E0,1.1791703E0,3.8600998E1,5.705572E0,1.9293294E0,1.6539946E0,2.6173878E0,1.4507532E1,1.6694213E0,1.2051893E0,6.220481E0,1.1816597E0,1.1975055E0,1.1916642E0,1.3677212E1,3.1615925E1,1.9008927E0,5.277312E0,4.0998106E0,1.4552524E0,2.1484613E0,1.1960965E0,3.1255623E1,3.3442297E0,6.196135E0,1.436073E0,3.6932316E1,1.6686838E0,1.4392505E0,1.1781373E0,4.033723E0,2.72219E1,4.5221367E0,2.7093788E1,1.652342E0,1.6918877E0,4.7701926E0,8.90702E0,1.8912159E0,3.8143563E0,1.2608207E1,1.8993256E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"101","size_leaf_vector":"0"}},{"base_weights":[-2.4452709E-2,1.7958993E-1,-3.7521696E-1,-1.5097151E-2,7.223681E-1,-1.9751932E-2,-7.5600904E-1,5.5232507E-1,-2.8102466E-1,-7.8672695E-1,2.2108553E-1,-3.386052E-1,9.61036E-1,3.8996243E-1,-9.21027E-1,9.18268E-1,-3.3353752E-1,1.4181946E-1,-1.0819033E0,-4.8212343E-1,3.499705E-1,-4.4018152E-1,1.1246376E0,-1.0402629E0,2.455809E-1,4.831688E-1,-9.921702E-1,-6.5828556E-1,1.6070692E-1,-7.748175E-2,9.611387E-1,1.0725286E0,-4.2544433E-1,1.03539545E-2,-4.1072956E-1,-1.7461775E-1,-1.2243621E0,1.2412341E0,-7.0094004E-2,9.047323E-1,-4.2246106E-1,-4.9621272E-1,6.088028E-1,-8.8695663E-1,2.6357803E-1,-3.4395626E-1,-7.333857E-2,-1.2196481E-1,1.2853296E0,-8.910565E-2,2.8627098E-1,-3.0939132E-1,1.6494498E-1,4.2910066E-1,-1.1249269E0,-2.5854045E-1,5.4370414E-2,1.1986954E0,-1.3513921E-1,7.668647E-1,-3.056961E-1,6.5772074E-1,-7.5581586E-1,-1.4802611E-1,3.011466E-1,-1.04118444E-1,-3.7046355E-1,1.4096668E0,7.148498E-2,-2.9040343E-1,2.7459365E-1,8.548135E-1,-7.814041E-1,-1.2747383E0,-1.5942918E-1,-1.1318598E-1,1.1425041E0,3.6626703E-1,-4.15371E-2,8.75088E-2,-3.9229384E-1,2.4193991E-2,4.2228568E-1,8.163276E-2,-2.7099752E-1,-1.3293221E0,2.8142953E-2,4.2116094E-1,-1.54368E-3,-7.92015E-2,3.1638458E-1,-3.2122314E-2,3.668394E-1,-4.202177E-1,2.8679892E-2,2.069931E-1,-2.0656578E-1,4.2017338E-1,-1.7089234E-1,1.4612564E-1,-2.84621E-1,-1.7977133E-1,3.258859E-1,9.252038E-2,-4.1487673E-1,-2.295309E-1,2.4182765E-1,-4.5870714E-2,-2.913981E-1,4.468622E-1,7.98479E-2,3.256611E-1,-1.1343567E-2,2.1128081E-1,-3.180195E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,true,true,true,false,false,true,true,true,true,true,true,true,false,false,true,false,false,true,true,true,false,false,true,true,false,false,true,true,true,false,false,false,true,true,false,true,true,true,true,true,true,false,false,true,false,false,false,false,false,true,false,false,false,true,false,true,false,true,false,false,false,false,false,true,true,false,false,false,true,false,true,true,true,false,false,true,false,false,false,false,false,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":2,"left_children":[1,3,5,7,11,9,13,15,27,17,29,25,21,19,23,31,53,-1,33,55,-1,51,47,35,-1,49,45,43,41,39,37,57,83,-1,-1,59,85,81,-1,77,61,79,75,73,71,-1,-1,69,67,-1,-1,-1,-1,63,65,-1,-1,87,-1,111,-1,89,113,-1,-1,-1,-1,109,105,-1,-1,91,107,103,101,99,97,-1,-1,95,-1,-1,-1,-1,-1,93,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[1.5955396E1,1.4913229E1,1.1079767E1,1.5964936E1,9.757921E0,8.194197E0,7.8375664E0,1.1254123E1,1.2104929E1,4.4503894E0,7.5480824E0,4.5505495E0,7.525894E0,4.4095263E0,8.0836525E0,5.4653473E0,6.8839025E0,0E0,2.8969011E0,1.0083029E0,0E0,2.8641267E0,5.7145805E0,5.056347E0,0E0,1.7647038E0,2.4479532E-1,8.461798E0,1.0301958E1,8.862179E0,3.5688934E0,4.7597656E0,1.314024E0,0E0,0E0,6.319068E0,3.874649E0,1.4497623E0,0E0,2.4194856E0,7.41173E0,6.900257E0,8.084208E0,8.684496E0,5.937142E0,0E0,0E0,4.4677863E0,3.6215858E0,0E0,0E0,0E0,0E0,3.742735E0,4.0198803E-2,0E0,0E0,5.083988E0,0E0,1.0319257E0,0E0,2.1307197E0,7.471982E0,0E0,0E0,0E0,0E0,1.6420288E0,2.6724727E0,0E0,0E0,2.227921E0,2.816987E-1,3.9215965E0,7.278205E0,5.319368E0,6.0825653E0,0E0,0E0,4.809184E0,0E0,0E0,0E0,0E0,0E0,2.6948853E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,13,13,12,12,14,14,11,11,8,8,10,10,15,15,18,18,23,23,30,30,29,29,28,28,27,27,26,26,22,22,25,25,21,21,16,16,19,19,31,31,35,35,40,40,53,53,54,54,48,48,47,47,44,44,43,43,42,42,39,39,41,41,37,37,32,32,36,36,57,57,61,61,71,71,85,85,79,79,76,76,75,75,74,74,73,73,68,68,72,72,67,67,59,59,62,62],"right_children":[2,4,6,8,12,10,14,16,28,18,30,26,22,20,24,32,54,-1,34,56,-1,52,48,36,-1,50,46,44,42,40,38,58,84,-1,-1,60,86,82,-1,78,62,80,76,74,72,-1,-1,70,68,-1,-1,-1,-1,64,66,-1,-1,88,-1,112,-1,90,114,-1,-1,-1,-1,110,106,-1,-1,92,108,104,102,100,98,-1,-1,96,-1,-1,-1,-1,-1,94,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[2.7524998E0,5.923786E0,1.4000399E0,-3.9256954E0,-1.0607927E0,-5.9424553E0,-6.8828187E0,4.677273E0,-1.5040994E-1,-8.633011E0,6.945169E-1,-1.536816E-1,-1.2275484E1,1.6851655E-1,9.9004E0,1.4918826E0,-1.753634E0,1.4181946E-1,-5.765919E-1,6.816157E-2,3.499705E-1,6.5075725E-1,-9.284999E0,-5.589089E0,2.455809E-1,1.8295392E-2,2.187073E-1,7.330673E0,-3.8353946E0,-5.0625286E0,9.9209505E-1,1.586516E0,2.5965471E0,1.03539545E-2,-4.1072956E-1,8.9877844E-1,1.1295551E1,-1.065483E0,-7.0094004E-2,4.250776E-1,-4.936661E0,3.377469E0,-1.4901059E0,2.9309592E0,1.5266433E-1,-3.4395626E-1,-7.333857E-2,-4.7487092E-1,1.2062815E1,-8.910565E-2,2.8627098E-1,-3.0939132E-1,1.6494498E-1,-2.442019E0,-5.5867515E0,-2.5854045E-1,5.4370414E-2,9.6940815E-1,-1.3513921E-1,-3.8390636E-1,-3.056961E-1,-1.9213891E0,-1.0100747E0,-1.4802611E-1,3.011466E-1,-1.04118444E-1,-3.7046355E-1,1.200358E1,1.8617029E0,-2.9040343E-1,2.7459365E-1,-7.724399E-1,-7.830738E-1,-1.6402028E0,1.8627863E0,1.15758814E-1,9.2337E-1,3.6626703E-1,-4.15371E-2,2.857921E-1,-3.9229384E-1,2.4193991E-2,4.2228568E-1,8.163276E-2,-2.7099752E-1,1.8974813E0,2.8142953E-2,4.2116094E-1,-1.54368E-3,-7.92015E-2,3.1638458E-1,-3.2122314E-2,3.668394E-1,-4.202177E-1,2.8679892E-2,2.069931E-1,-2.0656578E-1,4.2017338E-1,-1.7089234E-1,1.4612564E-1,-2.84621E-1,-1.7977133E-1,3.258859E-1,9.252038E-2,-4.1487673E-1,-2.295309E-1,2.4182765E-1,-4.5870714E-2,-2.913981E-1,4.468622E-1,7.98479E-2,3.256611E-1,-1.1343567E-2,2.1128081E-1,-3.180195E-1],"split_indices":[132,161,25,106,319,119,128,102,156,91,228,325,108,273,56,268,81,0,194,364,0,362,5,164,0,359,441,4,151,121,246,255,29,0,0,110,133,451,0,266,159,22,11,154,232,0,0,43,149,0,0,0,0,56,89,0,0,316,0,30,0,144,280,0,0,0,0,148,36,0,0,355,301,419,82,359,205,0,0,212,0,0,0,0,0,388,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[2.2098993E2,1.4002448E2,8.0965454E1,1.0377426E2,3.6250214E1,4.2419346E1,3.854611E1,3.2771393E1,7.100287E1,9.595315E0,3.282403E1,6.5801435E0,2.967007E1,4.6803246E0,3.3865784E1,2.3165375E1,9.606017E0,1.7076391E0,7.8876762E0,2.383253E0,2.2970715E0,2.925252E0,2.6744818E1,3.2069714E1,1.7960719E0,3.0304546E0,3.5496888E0,3.8031673E1,3.2971195E1,2.4026197E1,8.797834E0,2.0960848E1,2.2045279E0,1.7939317E0,6.0937443E0,5.9659495E0,2.6103764E1,7.09554E0,1.7022946E0,5.8234076E0,1.820279E1,1.3326635E1,1.964456E1,3.0503143E1,7.528533E0,2.5001907E0,1.049498E0,3.081377E0,2.366344E1,1.2799932E0,1.7504615E0,1.8054835E0,1.1197685E0,5.124998E0,4.481019E0,1.3254641E0,1.057789E0,1.955391E1,1.406936E0,2.864162E0,3.1017876E0,4.063299E0,1.413949E1,2.0215476E0,3.1034505E0,1.0697113E0,3.4113076E0,2.1318417E1,2.3450234E0,1.7339065E0,1.3474705E0,4.924376E0,2.6041567E0,1.9406872E1,1.1096271E1,8.684342E0,1.0960217E1,4.3382707E0,1.485137E0,8.269763E0,5.056872E0,1.0758486E0,6.0196915E0,1.0713257E0,1.1332022E0,2.4170704E1,1.9330604E0,1.6564594E1,2.989316E0,1.3238009E0,2.7394981E0,1.5587991E0,3.365577E0,2.3033466E1,1.137238E0,4.7180185E0,3.5517445E0,9.682706E0,1.2775112E0,5.2963896E0,3.3879528E0,8.584436E0,2.5118344E0,1.1229656E0,1.8283907E1,1.0769987E0,1.2680247E0,1.0232025E0,1.5809542E0,1.96437E1,1.6747179E0,1.8016344E0,1.0625275E0,2.2115462E0,1.1927944E1],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"115","size_leaf_vector":"0"}},{"base_weights":[-1.8392317E-2,-3.8326305E-1,1.2747033E-1,3.4788483E-1,-5.9278697E-1,3.8235024E-1,-1.541231E-1,9.7851294E-1,-1.1985575E-1,6.1970368E-2,7.1153796E-1,-6.8058366E-1,2.526655E-1,-4.4044074E-1,3.8460192E-1,5.0048355E-2,3.3774766E-1,6.1941236E-1,-2.9049554E-1,-6.7713517E-1,2.1875349E-1,1.0723836E0,-1.6146031E-1,-1.52486665E-2,-8.9678615E-1,-5.640963E-1,7.507844E-1,-3.6446917E-1,9.2586017E-1,-3.2018563E-1,1.1097592E0,-1.1032362E0,-1.9290815E-1,5.4957765E-1,-8.514013E-1,1.1916615E0,3.5432085E-1,6.669342E-1,-3.0102602E-1,7.2983995E-2,2.6339865E-1,-9.057904E-1,1.4892662E-1,6.012091E-1,-7.1486634E-1,-1.0141985E0,7.9892956E-2,4.0684482E-1,-1.2820873E-1,-5.889252E-1,4.6906522E-1,7.37072E-1,-4.6926957E-1,-3.0437595E-1,4.6507502E-1,9.718411E-2,-1.2471277E0,-1.2039482E-3,1.2628436E0,-1.6072133E-1,1.0253287E-1,1.9288743E-1,-9.0223247E-1,3.224095E-1,-1.8188128E-1,-1.8620016E-1,9.188155E-1,2.1135607E-1,-2.4592082E-1,-1.1110499E0,-6.1886903E-2,-1.1260736E0,9.967402E-2,-1.0591116E0,2.856863E-1,3.3089948E-1,-1.3877952E-1,-3.218097E-1,-4.361293E-2,6.3397296E-2,2.3436546E-1,-3.2002622E-1,-4.660085E-3,6.5824014E-1,-2.2008851E-1,-3.5583937E-1,3.1667045E-1,-7.8262076E-2,3.8572654E-1,3.7580258E-1,-6.557106E-1,-6.927883E-2,2.6025504E-1,3.9744022E-1,8.046765E-2,2.9112157E-1,-1.5547752E-1,2.2798961E-1,-1.9322787E-1,1.7507876E-1,-2.7913505E-1,-2.5498408E-1,1.2443223E-1,2.9250622E-1,-9.243116E-2,2.12605E-1,-8.088789E-2,-3.8158327E-1,-7.8123086E-4,2.588975E-1,-2.343867E-1,-3.8929486E-1,1.8751778E-2,-3.6082584E-1,3.6446344E-2,3.5765782E-1,-5.353755E-2,-4.0201798E-1,-1.0294644E-1,-3.2882772E-2,-3.2571092E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,true,true,true,true,false,false,false,false,true,false,false,true,false,false,false,true,true,false,false,true,true,false,true,false,true,true,false,false,true,true,false,false,false,false,true,false,false,false,true,false,true,true,true,true,false,true,true,false,true,true,false,false,false,false,false,true,false,false,false,true,false,false,false,true,false,false,true,true,true,true,true,false,false,false,false,false,false,false,false,false,false,false,true,false,false,false,true,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":3,"left_children":[1,3,5,7,11,9,13,15,25,17,27,23,-1,19,21,-1,-1,29,49,31,51,47,43,33,45,41,39,37,35,53,87,55,89,85,81,57,83,79,-1,-1,-1,77,-1,75,73,71,67,-1,59,69,65,63,61,-1,91,-1,117,-1,93,-1,-1,-1,119,-1,-1,-1,115,-1,-1,111,95,113,109,107,105,-1,-1,-1,-1,-1,-1,-1,-1,103,-1,101,-1,-1,-1,97,99,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[1.074657E1,9.03866E0,1.0419279E1,4.091213E0,6.265788E0,7.9771748E0,1.0857704E1,5.112324E-1,3.7598255E0,7.968201E0,8.966272E0,6.190651E0,0E0,7.280555E0,9.405109E0,0E0,0E0,7.5768137E0,5.9352136E0,6.7775173E0,4.9223733E0,3.924735E0,6.5798664E0,5.9442635E0,3.8120518E0,2.7263143E0,2.2491217E-3,5.2378197E0,4.360773E0,3.8258667E0,2.873867E0,4.061346E0,4.7392583E0,3.8245869E0,8.955579E-1,1.7981033E0,4.258071E0,4.266739E-3,0E0,0E0,0E0,4.9206924E-1,0E0,4.0709553E0,3.3683863E0,3.7406769E0,3.0604546E0,0E0,7.612166E-1,4.985321E0,4.2944083E0,4.14761E0,3.529563E0,0E0,1.2009399E0,0E0,9.165859E-1,0E0,8.7013626E-1,0E0,0E0,0E0,5.8263445E-1,0E0,0E0,0E0,1.9471459E0,0E0,0E0,2.088604E0,5.30786E0,2.6850853E0,3.0791068E0,1.5335817E0,9.1128343E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,3.1936893E0,0E0,1.6649585E0,0E0,0E0,0E0,3.739273E0,3.871482E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,13,13,14,14,11,11,8,8,10,10,17,17,19,19,23,23,28,28,27,27,26,26,25,25,22,22,24,24,21,21,18,18,20,20,29,29,31,31,35,35,48,48,52,52,51,51,50,50,46,46,49,49,45,45,44,44,43,43,41,41,37,37,34,34,36,36,33,33,30,30,32,32,54,54,58,58,70,70,89,89,90,90,85,85,83,83,74,74,73,73,72,72,69,69,71,71,66,66,56,56,62,62],"right_children":[2,4,6,8,12,10,14,16,26,18,28,24,-1,20,22,-1,-1,30,50,32,52,48,44,34,46,42,40,38,36,54,88,56,90,86,82,58,84,80,-1,-1,-1,78,-1,76,74,72,68,-1,60,70,66,64,62,-1,92,-1,118,-1,94,-1,-1,-1,120,-1,-1,-1,116,-1,-1,112,96,114,110,108,106,-1,-1,-1,-1,-1,-1,-1,-1,104,-1,102,-1,-1,-1,98,100,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[-4.469984E0,-5.237939E0,1.6427827E-1,-2.0697866E-1,1.239204E1,4.127674E-1,3.4596634E0,-8.466694E-1,3.3360985E-1,-5.0174356E-2,-7.411206E0,-5.4871025E0,2.526655E-1,4.499092E0,-2.1367402E0,5.0048355E-2,3.3774766E-1,-1.8120604E0,3.9632735E0,-9.7405887E-1,1.0555735E0,8.067951E-1,-3.3736491E0,-3.9070594E-1,7.8014493E0,7.9960597E-1,-3.1795868E-1,-4.6611834E0,2.9758868E0,-3.0011827E-1,-8.614849E0,-1.1712633E0,-3.2652023E0,-3.1230173E-1,9.93497E-1,-1.730041E0,6.893334E0,2.590583E-1,-3.0102602E-1,7.2983995E-2,2.6339865E-1,4.250776E-1,1.4892662E-1,7.662881E-1,3.9791207E0,1.160487E0,5.4196E0,4.0684482E-1,-3.420123E-1,-2.4716605E-1,-4.8323867E-1,1.1477818E0,-8.042161E0,-3.0437595E-1,-2.7074695E-1,9.718411E-2,1.3423651E0,-1.2039482E-3,1.2188125E0,-1.6072133E-1,1.0253287E-1,1.9288743E-1,-7.94084E0,3.224095E-1,-1.8188128E-1,-1.8620016E-1,5.065042E-1,2.1135607E-1,-2.4592082E-1,4.8222556E0,-7.177088E-1,1.6130679E0,-1.1262336E-1,3.5519876E0,8.291802E0,3.3089948E-1,-1.3877952E-1,-3.218097E-1,-4.361293E-2,6.3397296E-2,2.3436546E-1,-3.2002622E-1,-4.660085E-3,2.3108504E0,-2.2008851E-1,2.1182984E-1,3.1667045E-1,-7.8262076E-2,3.8572654E-1,1.2814288E0,-9.075556E-1,-6.927883E-2,2.6025504E-1,3.9744022E-1,8.046765E-2,2.9112157E-1,-1.5547752E-1,2.2798961E-1,-1.9322787E-1,1.7507876E-1,-2.7913505E-1,-2.5498408E-1,1.2443223E-1,2.9250622E-1,-9.243116E-2,2.12605E-1,-8.088789E-2,-3.8158327E-1,-7.8123086E-4,2.588975E-1,-2.343867E-1,-3.8929486E-1,1.8751778E-2,-3.6082584E-1,3.6446344E-2,3.5765782E-1,-5.353755E-2,-4.0201798E-1,-1.0294644E-1,-3.2882772E-2,-3.2571092E-1],"split_indices":[154,121,147,237,56,16,36,260,447,25,3,8,0,5,37,0,0,76,45,151,69,403,64,40,17,369,190,141,109,178,2,187,26,480,373,292,117,386,0,0,0,266,0,194,97,284,131,0,271,175,257,413,67,0,44,0,360,0,342,0,0,0,126,0,0,0,483,0,0,89,212,197,370,115,97,0,0,0,0,0,0,0,0,113,0,317,0,0,0,57,441,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[2.000055E2,5.6655647E1,1.4334984E2,1.2472313E1,4.4183334E1,7.505614E1,6.82937E1,4.846309E0,7.626004E0,3.8639114E1,3.6417023E1,4.2131374E1,2.0519595E0,4.471261E1,2.3581093E1,1.0720179E0,3.7742913E0,1.467404E1,2.3965075E1,3.287948E1,1.183313E1,1.0009711E1,1.3571382E1,1.0595536E1,3.153584E1,5.282646E0,2.3433578E0,5.932463E0,3.0484558E1,5.1509194E0,9.523121E0,1.680946E1,1.6070019E1,6.5283356E0,4.0672007E0,2.0067106E1,1.0417453E1,2.2525578E0,3.6799054E0,1.053373E0,1.2899848E0,4.1057053E0,1.1769408E0,5.6704054E0,7.900976E0,2.811728E1,3.4185605E0,7.991351E0,2.0183601E0,1.7366476E1,6.598598E0,6.7082305E0,5.1248994E0,2.5786226E0,2.5722969E0,1.4249372E0,1.5384522E1,1.180551E0,1.8886555E1,1.0060763E0,1.0122838E0,1.3000962E0,3.8248034E0,5.5206833E0,1.1875472E0,1.816717E0,4.781881E0,2.1456661E0,1.2728944E0,8.167259E0,9.199217E0,2.5541037E1,2.5762432E0,5.8349166E0,2.0660596E0,3.827952E0,1.8424535E0,3.0299604E0,1.0757449E0,1.0230669E0,1.2294909E0,3.014359E0,1.0528418E0,8.451494E0,1.9659587E0,2.4467964E0,4.081539E0,1.0287875E0,8.494333E0,7.2943745E0,8.775644E0,1.0985626E0,1.4737343E0,1.7466553E1,1.4200017E0,2.4748552E0,6.7243614E0,5.4836965E0,1.8106778E0,1.3884389E0,7.387205E0,1.4281418E0,1.0186546E0,6.358209E0,2.093285E0,1.0159751E0,1.0500845E0,4.686842E0,1.1480745E0,1.391214E0,1.1850291E0,6.955559E0,1.2117007E0,2.409594E1,1.4450963E0,3.758881E0,1.0229998E0,1.3512926E1,1.871596E0,1.0191838E0,2.8056197E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"121","size_leaf_vector":"0"}},{"base_weights":[-1.1468058E-2,-1.4141731E-1,3.2650363E-1,-3.7894976E-1,1.7035E-1,-9.2376605E-2,5.6921864E-1,-6.739002E-1,-1.1986341E-1,3.2334623E-1,-6.455749E-1,3.8112214E-1,-5.902798E-1,7.0633763E-1,-4.28209E-1,-8.171325E-1,3.2762265E-1,-2.1145584E-1,6.2683076E-1,1.0436684E0,3.4273058E-1,-8.6925006E-1,2.0116186E-1,-2.8946915E-1,1.290125E-1,-8.088682E-2,7.752188E-1,2.2442168E-1,-5.3519905E-1,1.6875917E-1,-9.346314E-1,1.7849615E-1,-9.10311E-1,8.7391084E-1,-1.7732129E-1,2.600256E-1,-6.8515176E-1,2.49213E-2,-3.39238E-1,5.566201E-1,-7.236398E-1,6.286255E-1,-2.9538068E-1,-1.43698165E-2,-3.4505096E-1,9.0123236E-1,-2.3936836E-1,-5.0476694E-1,7.2784656E-1,2.7104858E-1,-1.3126367E-1,1.1645178E0,-2.8613072E-2,7.004788E-2,-9.9224854E-1,9.0891635E-1,-1.8473428E-1,8.5721713E-1,-1.5517728E-1,3.677575E-1,6.797741E-2,9.505039E-1,-1.2985614E-1,-2.8895354E-1,9.3767814E-2,1.0020688E0,-6.60191E-2,-2.0019867E-1,1.5769975E-1,-9.228643E-1,3.9467517E-1,3.6204025E-1,-7.567555E-1,7.355999E-3,2.335899E-1,-7.922564E-3,3.3021754E-1,-9.745201E-1,1.3188595E-1,-3.1684065E-1,7.909158E-2,-2.0866874E-1,1.7674142E-1,-3.1308752E-1,3.2666575E-2,-7.8100204E-3,-3.4379742E-1,-2.94993E-1,5.0569803E-2,2.4658383E-1,-2.0028456E-1,3.2374424E-1,-6.367391E-2,2.316487E-1,-6.373709E-2,3.2645816E-1,-2.3283057E-2,3.4292293E-1,-8.084051E-3,-9.2479765E-2,3.1487933E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,true,true,true,true,true,false,true,true,true,true,true,true,true,false,true,true,true,true,false,false,false,true,true,false,true,false,true,false,true,true,false,false,true,false,false,false,true,true,true,true,false,false,false,true,true,false,false,true,false,false,true,true,false,true,false,false,false,true,false,false,false,true,false,false,false,true,true,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":4,"left_children":[1,3,5,7,11,9,13,15,27,17,29,25,21,19,23,31,49,-1,33,51,47,43,-1,-1,-1,35,45,41,39,-1,37,-1,53,75,-1,55,77,-1,-1,73,69,57,71,67,-1,65,-1,63,61,-1,-1,59,-1,-1,79,97,81,99,-1,-1,-1,95,-1,-1,-1,91,-1,-1,-1,83,93,89,87,-1,-1,-1,-1,85,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[7.9283085E0,9.715784E0,5.15062E0,5.5954657E0,9.370589E0,4.6372457E0,4.6635876E0,5.219494E0,5.948551E0,4.080395E0,3.5274892E0,8.306408E0,5.0744977E0,3.127081E0,2.4768453E0,4.552984E0,2.6129942E0,0E0,3.3850327E0,2.0724955E0,5.3160214E0,2.5992012E0,0E0,0E0,0E0,4.6582274E0,5.581436E0,5.016198E0,4.2644396E0,0E0,1.5536442E0,0E0,2.9490528E0,1.7297258E0,0E0,4.3897133E0,3.0710096E0,0E0,0E0,3.740387E-1,4.0273037E0,3.9149446E0,3.5308776E0,1.6578093E0,0E0,2.8412895E0,0E0,2.2971768E0,3.2114224E0,0E0,0E0,4.6774292E-1,0E0,0E0,2.464571E0,1.3715053E0,4.1694603E0,2.885315E0,0E0,0E0,0E0,1.4156008E0,0E0,0E0,0E0,2.2363873E0,0E0,0E0,0E0,1.8897238E0,8.9435697E-1,2.965918E0,1.556303E0,0E0,0E0,0E0,0E0,1.0710154E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,13,13,12,12,14,14,11,11,8,8,10,10,15,15,18,18,25,25,30,30,28,28,27,27,21,21,26,26,20,20,16,16,19,19,32,32,35,35,41,41,51,51,48,48,47,47,45,45,43,43,40,40,42,42,39,39,33,33,36,36,54,54,56,56,69,69,77,77,72,72,71,71,65,65,70,70,61,61,55,55,57,57],"right_children":[2,4,6,8,12,10,14,16,28,18,30,26,22,20,24,32,50,-1,34,52,48,44,-1,-1,-1,36,46,42,40,-1,38,-1,54,76,-1,56,78,-1,-1,74,70,58,72,68,-1,66,-1,64,62,-1,-1,60,-1,-1,80,98,82,100,-1,-1,-1,96,-1,-1,-1,92,-1,-1,-1,84,94,90,88,-1,-1,-1,-1,86,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[4.8019023E0,1.551867E0,-1.273237E0,-2.920134E-1,7.1497593E0,-7.980547E-2,1.2979289E0,1.0874666E1,2.3395972E0,-5.530925E0,-6.0904665E0,-2.9370213E-1,6.8734803E0,-1.1687138E0,-1.9992657E0,-1.4647161E1,-2.4839075E-1,-2.1145584E-1,1.1660391E0,9.5685047E-1,-2.0337958E0,-2.5694861E0,2.0116186E-1,-2.8946915E-1,1.290125E-1,4.0720785E-1,1.185528E1,-1.846885E-2,-6.802229E0,1.6875917E-1,-8.5699356E-1,1.7849615E-1,-1.1771767E1,-8.272495E-1,-1.7732129E-1,-3.884367E0,6.604327E-1,2.49213E-2,-3.39238E-1,-3.021791E0,1.2929546E0,5.791294E0,-1.4279E-1,1.3024313E0,-3.4505096E-1,1.4857714E0,-2.3936836E-1,2.7173644E-1,6.0546117E0,2.7104858E-1,-1.3126367E-1,4.575576E0,-2.8613072E-2,7.004788E-2,1.8086476E0,2.8110227E-1,4.4334164E-1,-7.596429E0,-1.5517728E-1,3.677575E-1,6.797741E-2,1.1608107E0,-1.2985614E-1,-2.8895354E-1,9.3767814E-2,1.6648287E0,-6.60191E-2,-2.0019867E-1,1.5769975E-1,1.2193571E0,-5.368619E-1,5.838567E0,5.343601E-1,7.355999E-3,2.335899E-1,-7.922564E-3,3.3021754E-1,-4.177292E0,1.3188595E-1,-3.1684065E-1,7.909158E-2,-2.0866874E-1,1.7674142E-1,-3.1308752E-1,3.2666575E-2,-7.8100204E-3,-3.4379742E-1,-2.94993E-1,5.0569803E-2,2.4658383E-1,-2.0028456E-1,3.2374424E-1,-6.367391E-2,2.316487E-1,-6.373709E-2,3.2645816E-1,-2.3283057E-2,3.4292293E-1,-8.084051E-3,-9.2479765E-2,3.1487933E-1],"split_indices":[70,3,161,11,110,372,261,161,81,45,121,78,94,61,49,30,410,0,222,233,28,61,0,0,0,224,18,272,49,0,168,0,144,356,0,12,418,0,0,129,392,68,473,7,0,496,0,264,10,0,0,104,0,0,253,368,383,80,0,0,0,210,0,0,0,263,0,0,0,488,327,90,429,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.784691E2,1.293254E2,4.914371E1,7.3279205E1,5.6046192E1,1.825626E1,3.0887451E1,3.3519123E1,3.9760082E1,1.06605015E1,7.5957594E0,4.4276123E1,1.177007E1,2.7411026E1,3.4764247E0,2.9486269E1,4.0328546E0,2.12808E0,8.532421E0,1.3249082E1,1.4161943E1,9.891936E0,1.8781338E0,2.062454E0,1.4139707E0,2.0747768E1,2.3528355E1,2.1990833E1,1.7769249E1,1.2838485E0,6.3119106E0,1.4918507E0,2.799442E1,7.3281736E0,1.2042474E0,1.3628541E1,7.1192265E0,1.1131775E0,5.198733E0,2.3278646E0,1.5441384E1,1.2253234E1,9.737599E0,2.7065768E0,7.1853595E0,2.2178831E1,1.3495241E0,4.3276668E0,9.834277E0,2.1979084E0,1.834946E0,1.1958009E1,1.2910731E0,1.7477082E0,2.6246712E1,5.1179705E0,8.510571E0,1.0425393E1,1.8278407E0,1.0935817E1,1.0221924E0,8.403355E0,1.4309222E0,2.6582763E0,1.6693904E0,2.0446445E1,1.7323861E0,1.1184918E0,1.588085E0,1.3254845E1,2.1865394E0,4.104429E0,5.6331706E0,1.0122471E0,1.3156176E0,1.6524569E0,5.675717E0,5.7687044E0,1.3505223E0,2.5127527E1,1.1191835E0,5.1875167E0,3.323054E0,1.1855698E1,1.3991472E0,1.0595678E0,4.7091365E0,4.4748783E0,1.1582923E0,2.9766192E0,1.1278096E0,1.9330738E1,1.1157078E0,1.1806473E0,1.0058922E0,7.3593235E0,1.044031E0,3.9165947E0,1.2013756E0,1.3917142E0,9.033679E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"101","size_leaf_vector":"0"}},{"base_weights":[-9.289552E-3,4.920656E-1,-8.097947E-2,1.0629389E0,1.0588677E-1,-1.9874333E-1,2.8666547E-1,9.4871685E-2,3.374571E-1,-8.8609815E-2,-7.273655E-1,-3.058748E-1,7.693089E-1,7.100358E-1,-5.1558767E-2,-2.910857E-1,2.1148232E-1,-5.9626824E-1,1.6115855E-1,3.1479472E-1,-6.05081E-1,-1.2306485E-1,8.644954E-1,9.020416E-3,-9.233638E-1,-2.0398702E-2,3.0136183E-1,-1.6422309E-1,-8.3288133E-1,-8.651106E-1,5.242748E-2,6.204619E-1,-2.2828966E-1,-8.685931E-2,9.886178E-1,-1.0821581E0,8.585726E-2,9.396452E-2,-3.0195856E-1,4.297628E-1,-3.3288085E-1,6.285179E-1,-5.0788593E-1,-4.1652578E-1,2.1253853E-1,-3.064891E-1,-3.347849E-2,-1.5637715E-1,2.2976083E-1,-2.5525454E-1,7.279176E-2,8.6966574E-1,-1.1714693E-1,-5.914707E-1,5.0447184E-1,-3.4442228E-1,-5.228178E-2,6.917873E-1,-9.869501E-2,1.3319002E-2,1.0690204E0,-1.0208306E0,6.2250566E-2,5.4641847E-2,2.2498685E-1,-1.4949388E-2,-2.5191912E-1,-1.7908303E-2,3.07975E-1,1.6981289E-1,-1.5512033E-1,-3.3690923E-1,-4.102896E-2,2.7565393E-1,-3.077946E-3,3.437647E-1,8.84236E-2,2.1519932E-1,1.0383202E-2,-6.3630775E-2,1.9692549E-1,1.5373439E-1,-2.4240129E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,false,true,true,false,false,true,true,true,true,false,true,true,true,true,false,true,false,false,false,true,false,false,false,false,true,true,false,false,false,false,false,true,false,true,false,true,false,true,true,true,true,false,false,false,false,false,false,false,false,true,true,false,false,true,true,false,true,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":5,"left_children":[1,3,5,7,11,9,13,-1,-1,15,23,17,25,21,19,27,39,29,-1,41,37,-1,33,31,35,-1,-1,43,61,45,-1,63,-1,-1,59,55,-1,47,-1,57,53,51,49,65,79,-1,-1,-1,-1,-1,-1,67,-1,81,77,-1,-1,73,69,-1,75,71,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[5.7935934E0,4.3337445E0,6.157944E0,1.541996E-2,3.8250287E0,6.183544E0,4.990635E0,0E0,0E0,5.527276E0,2.658966E0,2.4805846E0,1.1155436E0,3.0248919E0,4.2795377E0,3.5840297E0,4.5181537E0,1.6098714E0,0E0,3.6206274E0,2.4443011E0,0E0,2.2749748E0,2.755951E0,3.0974474E0,0E0,0E0,4.3764706E0,2.172113E0,5.2119684E-1,0E0,4.8138738E-2,0E0,0E0,9.214144E-1,5.8023167E-1,0E0,2.0398827E0,0E0,3.7678905E0,2.7023008E0,2.7249413E0,1.2577778E0,4.2032185E0,3.674193E0,0E0,0E0,0E0,0E0,0E0,0E0,1.2511921E0,0E0,2.5516934E0,2.9114014E-1,0E0,0E0,2.8765926E0,2.9990873E0,0E0,3.595829E-1,6.2885857E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,9,9,11,11,14,14,13,13,10,10,12,12,15,15,17,17,23,23,22,22,24,24,20,20,16,16,19,19,27,27,29,29,37,37,42,42,41,41,40,40,35,35,39,39,34,34,28,28,31,31,43,43,51,51,58,58,61,61,57,57,60,60,54,54,44,44,53,53],"right_children":[2,4,6,8,12,10,14,-1,-1,16,24,18,26,22,20,28,40,30,-1,42,38,-1,34,32,36,-1,-1,44,62,46,-1,64,-1,-1,60,56,-1,48,-1,58,54,52,50,66,80,-1,-1,-1,-1,-1,-1,68,-1,82,78,-1,-1,74,70,-1,76,72,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[-8.258678E0,-2.5611613E0,5.0261116E0,-9.498693E0,3.2307458E0,7.420925E0,-1.7640333E0,9.4871685E-2,3.374571E-1,3.26769E0,-5.863694E0,1.1113319E0,-2.0008628E-1,-1.2291954E0,2.0557225E0,7.2363443E0,6.1032486E-1,3.0782157E-1,1.6115855E-1,5.4211454E0,-3.8969758E-1,-1.2306485E-1,-8.478059E0,2.8290343E-1,1.0473132E1,-2.0398702E-2,3.0136183E-1,1.3046451E0,9.93506E-1,7.24233E-1,5.242748E-2,-5.3265963E-2,-2.2828966E-1,-8.685931E-2,-1.4108841E0,1.0584699E0,8.585726E-2,1.2116785E0,-3.0195856E-1,4.9717855E-1,1.0629382E0,7.3260574E0,4.3602604E-1,7.5094607E-3,-1.0717907E0,-3.064891E-1,-3.347849E-2,-1.5637715E-1,2.2976083E-1,-2.5525454E-1,7.279176E-2,-1.4687598E0,-1.1714693E-1,-7.43776E-1,-8.82411E-2,-3.4442228E-1,-5.228178E-2,5.7521474E-1,-2.977004E0,1.3319002E-2,1.3367885E0,9.250545E-1,6.2250566E-2,5.4641847E-2,2.2498685E-1,-1.4949388E-2,-2.5191912E-1,-1.7908303E-2,3.07975E-1,1.6981289E-1,-1.5512033E-1,-3.3690923E-1,-4.102896E-2,2.7565393E-1,-3.077946E-3,3.437647E-1,8.84236E-2,2.1519932E-1,1.0383202E-2,-6.3630775E-2,1.9692549E-1,1.5373439E-1,-2.4240129E-1],"split_indices":[159,134,5,63,111,81,64,0,0,36,57,404,204,264,69,75,189,278,0,117,232,0,77,157,90,0,0,73,169,381,0,442,0,0,424,225,0,47,0,280,401,123,230,226,48,0,0,0,0,0,0,264,0,310,103,0,0,317,64,0,338,302,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.5908371E2,1.9168339E1,1.3991537E2,7.027461E0,1.2140878E1,1.063188E2,3.3596577E1,1.0057708E0,6.0216904E0,8.895374E1,1.7365059E1,7.8238487E0,4.317029E0,1.4432041E1,1.9164536E1,5.3133755E1,3.581999E1,6.037119E0,1.7867297E0,1.1791511E1,7.373025E0,1.5481225E0,1.2883919E1,3.8508809E0,1.3514178E1,1.1027437E0,3.2142854E0,4.4108715E1,9.02504E0,4.3858366E0,1.6512825E0,2.2546594E0,1.5962214E0,1.1217585E0,1.176216E1,1.2038986E1,1.4751922E0,2.9690015E0,4.4040236E0,2.5718018E1,1.0101972E1,8.707125E0,3.0843854E0,2.635455E1,1.7754166E1,3.3368833E0,1.0489532E0,1.6669178E0,1.3020837E0,2.02525E0,1.0591356E0,7.1594095E0,1.5477154E0,7.9427476E0,2.1592243E0,1.0977423E1,1.0615637E0,1.6986538E1,8.73148E0,1.0447814E0,1.0717379E1,7.660021E0,1.3650188E0,1.0630181E0,1.1916413E0,1.4718844E1,1.1635705E1,1.1237094E0,6.0357E0,3.236366E0,5.495114E0,6.623975E0,1.0360459E0,1.2601951E1,4.3845873E0,9.312946E0,1.4044319E0,1.1235433E0,1.035681E0,9.340584E0,8.413582E0,1.0783857E0,6.864362E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"83","size_leaf_vector":"0"}},{"base_weights":[-5.7448456E-3,-1.5914804E-1,2.1152855E-1,-2.3314597E-1,6.198741E-1,-1.6293481E-3,6.385403E-1,-7.56179E-2,-6.698242E-1,-1.7910734E-1,6.212127E-1,-1.6620778E-1,8.9096165E-1,-1.24407075E-1,8.341805E-1,5.853577E-2,-6.4914274E-1,-5.8930784E-1,1.35365E-1,4.626424E-1,-2.3805787E-1,-5.780606E-2,9.5123065E-1,3.190736E-1,-1.2840627E-2,-8.5113865E-1,1.5169694E-1,-1.0300866E-1,8.354343E-1,2.864975E-1,-3.728646E-1,-8.0476165E-1,2.4359745E-1,-5.974684E-2,-9.7790134E-1,2.963554E-1,-2.7278135E-2,1.0430447E0,-1.2809567E-2,-1.6899316E-1,2.2035812E-1,-1.7638652E-2,2.2496344E-1,-9.1077524E-1,1.5397052E-1,4.8180524E-1,-3.397816E-1,7.3732173E-1,5.2313966E-4,-1.1426938E-1,-3.2284397E-1,3.0986479E-2,3.3533475E-1,-6.194103E-1,3.7095708E-1,4.2891647E-3,8.823596E-1,-1.6380996E-1,1.987041E-1,-1.0653656E0,-6.752401E-3,1.5284392E-2,-3.1500953E-1,-1.8811043E-1,1.326007E-1,4.524678E-1,-5.688925E-1,1.9855021E-1,-1.2487322E-1,-2.073633E-3,3.0362514E-1,-1.5867567E-1,1.9253752E-1,2.9838726E-1,4.677801E-2,6.411086E-2,-2.2704987E-1,2.555313E-1,-9.458662E-2,2.1962427E-1,-1.0858504E-1,-9.640269E-2,-3.3998406E-1,1.9303547E-1,-2.675283E-2,1.3741872E-1,-1.0116222E-1,3.4262046E-2,-2.833008E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false,false,true,false,true,false,true,false,true,false,true,true,true,false,true,false,false,false,true,false,false,false,true,false,true,true,false,false,true,true,true,false,false,true,false,false,false,false,false,false,false,true,true,false,true,true,false,false,false,true,false,false,false,false,false,false,false,false,false,false,false,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":6,"left_children":[1,3,5,7,11,9,13,15,25,17,27,-1,23,19,21,29,43,31,45,41,-1,-1,37,-1,-1,33,39,-1,35,47,65,49,67,63,59,-1,-1,51,-1,-1,-1,-1,-1,61,57,55,53,69,85,71,-1,-1,-1,87,83,79,73,-1,-1,81,-1,-1,-1,-1,-1,77,75,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[4.890237E0,5.100869E0,5.52882E0,5.400688E0,2.9541166E0,4.6994467E0,3.1224985E0,4.6056037E0,3.23952E0,4.346781E0,2.2350469E0,0E0,1.1197653E0,2.3302877E0,2.1336145E0,4.928464E0,2.5222745E0,2.8054943E0,3.3797297E0,5.374307E-1,0E0,0E0,1.3891535E0,0E0,0E0,1.6728296E0,2.3407917E0,0E0,1.2202373E0,4.208164E0,3.070143E0,2.0854335E0,1.3120489E0,1.2584879E0,1.2135706E0,0E0,0E0,8.3467484E-1,0E0,0E0,0E0,0E0,0E0,1.2180533E0,1.6573845E0,2.2364774E0,1.9836285E0,2.597703E0,3.3955188E0,1.7326086E0,0E0,0E0,0E0,1.7067587E0,4.4821224E-1,1.9569671E0,3.639655E-1,0E0,0E0,3.2031727E-1,0E0,0E0,0E0,0E0,0E0,1.4622444E0,2.2931776E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,13,13,14,14,12,12,8,8,10,10,15,15,17,17,25,25,28,28,22,22,26,26,19,19,16,16,18,18,29,29,31,31,37,37,46,46,45,45,44,44,34,34,43,43,33,33,30,30,32,32,47,47,49,49,56,56,66,66,65,65,55,55,59,59,54,54,48,48,53,53],"right_children":[2,4,6,8,12,10,14,16,26,18,28,-1,24,20,22,30,44,32,46,42,-1,-1,38,-1,-1,34,40,-1,36,48,66,50,68,64,60,-1,-1,52,-1,-1,-1,-1,-1,62,58,56,54,70,86,72,-1,-1,-1,88,84,80,74,-1,-1,82,-1,-1,-1,-1,-1,78,76,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[1.7952664E0,1.0111593E1,4.3376073E-1,4.2677717E0,-9.723111E0,5.6921854E0,-1.1436375E0,6.8683324E0,9.2868316E-1,-2.4635348E0,-7.459593E0,-1.6620778E-1,5.4731503E0,4.731121E-1,-1.1049088E1,2.7524998E0,3.2010233E0,9.883626E-1,1.9237041E0,8.08115E-2,-2.3805787E-1,-5.780606E-2,1.1543075E0,3.190736E-1,-1.2840627E-2,-1.284073E0,-1.4288657E0,-1.0300866E-1,6.954727E0,-2.6673143E0,-8.346301E-1,-5.0363934E-1,4.8559037E-1,-2.7798134E-1,8.544391E0,2.963554E-1,-2.7278135E-2,-1.2060851E0,-1.2809567E-2,-1.6899316E-1,2.2035812E-1,-1.7638652E-2,2.2496344E-1,-6.415518E-1,4.7408724E0,-2.7061972E-1,5.4135394E-1,-3.2207503E0,-3.762381E0,5.3940934E-1,-3.2284397E-1,3.0986479E-2,3.3533475E-1,-4.0876076E-1,-1.6042028E-1,-4.7301745E0,8.087434E-1,-1.6380996E-1,1.987041E-1,-9.1293395E-1,-6.752401E-3,1.5284392E-2,-3.1500953E-1,-1.8811043E-1,1.326007E-1,1.487741E0,-7.4083745E-1,1.9855021E-1,-1.2487322E-1,-2.073633E-3,3.0362514E-1,-1.5867567E-1,1.9253752E-1,2.9838726E-1,4.677801E-2,6.411086E-2,-2.2704987E-1,2.555313E-1,-9.458662E-2,2.1962427E-1,-1.0858504E-1,-9.640269E-2,-3.3998406E-1,1.9303547E-1,-2.675283E-2,1.3741872E-1,-1.0116222E-1,3.4262046E-2,-2.833008E-1],"split_indices":[133,56,264,27,112,161,415,68,324,73,99,0,20,168,7,132,76,419,121,292,0,0,459,0,0,199,105,0,128,86,285,464,437,435,31,0,0,316,0,0,0,0,0,258,145,431,392,60,61,382,0,0,0,172,452,124,489,0,0,294,0,0,0,0,0,67,323,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.4470987E2,8.497901E1,5.9730854E1,7.824717E1,6.73184E0,4.05067E1,1.9224155E1,5.8365467E1,1.9881699E1,3.207843E1,8.42827E0,1.0671232E0,5.664717E0,3.9978077E0,1.5226347E1,4.802892E1,1.0336547E1,1.3541205E1,1.8537226E1,2.293869E0,1.7039387E0,1.491899E0,1.37344475E1,4.663578E0,1.0011392E0,1.6257519E1,3.6241808E0,1.4221092E0,7.006161E0,3.1643608E1,1.6385311E1,1.0785127E1,2.7560785E0,2.447689E0,1.380983E1,5.9592853E0,1.046876E0,1.2527781E1,1.2066674E0,1.6425796E0,1.9816012E0,1.0972552E0,1.1966138E0,7.6955037E0,2.641043E0,1.0712491E1,7.8247347E0,1.1669285E1,1.9974323E1,3.4438639E0,7.341263E0,1.098405E0,1.1429376E1,5.707649E0,2.1170852E0,5.3478384E0,5.3646526E0,1.0799476E0,1.5610954E0,1.254615E1,1.2636799E0,1.0535688E0,6.641935E0,1.0316203E0,1.4160688E0,2.893727E0,1.3491584E1,1.6850194E0,1.0710591E0,3.409363E0,8.259922E0,2.4181159E0,1.025748E0,4.326792E0,1.0378606E0,2.5860472E0,1.0905537E1,1.8316658E0,1.0620613E0,1.4558517E0,3.8919868E0,1.5288407E0,1.1017309E1,1.0794334E0,1.0376518E0,8.331269E0,1.1643054E1,1.9510596E0,3.7565897E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"89","size_leaf_vector":"0"}},{"base_weights":[-1.6243302E-3,7.488653E-2,-4.0100363E-1,-7.1237236E-1,1.2939765E-1,7.812229E-2,-7.530622E-1,1.2874591E-2,-8.85273E-1,6.7541283E-1,-2.5846353E-1,3.1503406E-1,-8.3547786E-2,4.9958117E-2,-8.7207687E-1,-2.9416013E-1,-8.0123484E-2,-6.15196E-1,1.3587965E-1,-9.9778587E-1,-5.1056392E-2,1.1183825E-1,-4.7631073E-1,2.4728505E-1,3.814301E-2,3.8855535E-1,-4.9581677E-1,-2.5261858E-1,2.6244245E-2,-6.739699E-1,2.8219652E-1,3.1805832E-2,-2.3477565E-1,-5.1406164E-2,5.6225723E-1,-3.1719583E-1,-9.943506E-2,-7.5730383E-1,1.8154609E-1,-2.756752E-1,6.321001E-2,3.4450972E-1,-4.7052595E-1,-3.6959368E-1,2.0787789E-1,4.9424604E-2,-9.414743E-1,6.820796E-1,5.6042224E-2,6.702125E-1,-1.6192329E-1,-9.568193E-2,1.9677755E-1,6.1706122E-2,-2.0972829E-1,9.860025E-2,2.6938275E-1,1.11095436E-1,-1.3142528E-1,1.4171496E-1,-1.9214366E-1,2.5480178E-1,-6.906645E-2,-2.3408173E-1,5.1842995E-2,-4.9602617E-2,-3.082546E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false,false,false,false,true,true,true,true,true,false,true,false,false,true,false,true,false,false,true,false,false,true,true,false,false,false,false,false,false,true,false,false,false,true,true,false,false,false,false,true,false,false,true,true,false,true,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":7,"left_children":[1,3,5,7,11,9,13,-1,15,23,17,25,21,-1,19,-1,-1,27,-1,35,-1,29,37,-1,-1,33,31,-1,-1,39,47,-1,-1,41,49,-1,-1,45,43,-1,-1,51,63,53,-1,-1,65,61,57,55,59,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[4.025365E0,4.8373117E0,3.6312664E0,9.9314046E-1,4.147684E0,2.160141E0,1.4916348E0,0E0,9.13353E-2,2.0833278E-1,2.0652595E0,3.5257325E0,3.8402364E0,0E0,8.054209E-1,0E0,0E0,8.8872683E-1,0E0,1.3251305E-2,0E0,4.7101464E0,3.20304E0,0E0,0E0,3.989318E0,9.609916E-1,0E0,0E0,1.5611525E0,2.5349507E0,0E0,0E0,2.7482274E0,3.019742E0,0E0,0E0,2.161242E0,1.8823444E0,0E0,0E0,1.9843165E0,1.7276868E0,7.9382205E-1,0E0,0E0,4.9761963E-1,1.7503071E0,3.1537912E0,2.240614E0,2.0134783E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,8,8,10,10,14,14,12,12,9,9,11,11,17,17,21,21,26,26,25,25,19,19,22,22,29,29,33,33,38,38,37,37,30,30,34,34,41,41,43,43,49,49,48,48,50,50,47,47,42,42,46,46],"right_children":[2,4,6,8,12,10,14,-1,16,24,18,26,22,-1,20,-1,-1,28,-1,36,-1,30,38,-1,-1,34,32,-1,-1,40,48,-1,-1,42,50,-1,-1,46,44,-1,-1,52,64,54,-1,-1,66,62,58,56,60,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[8.007231E0,-1.0792214E1,-1.7596493E0,-5.4872627E0,1.139224E0,-4.4178343E-1,-1.3470718E0,1.2874591E-2,6.433465E-1,-1.4874662E-2,4.0164137E-1,1.0730743E1,3.0934224E0,4.9958117E-2,2.7016306E0,-2.9416013E-1,-8.0123484E-2,5.781299E-1,1.3587965E-1,6.826568E0,-5.1056392E-2,-6.7912426E0,2.1094308E0,2.4728505E-1,3.814301E-2,-5.2722406E0,-2.4366713E-1,-2.5261858E-1,2.6244245E-2,7.309173E-1,-3.9422352E0,3.1805832E-2,-2.3477565E-1,-7.08375E-2,1.1170988E0,-3.1719583E-1,-9.943506E-2,-5.7656326E0,-6.855499E-2,-2.756752E-1,6.321001E-2,-4.562254E0,1.2233633E-1,-9.5463985E-1,2.0787789E-1,4.9424604E-2,-1.3531911E0,6.888371E0,1.8133402E-1,-2.0798903E0,3.9204042E-2,-9.568193E-2,1.9677755E-1,6.1706122E-2,-2.0972829E-1,9.860025E-2,2.6938275E-1,1.11095436E-1,-1.3142528E-1,1.4171496E-1,-1.9214366E-1,2.5480178E-1,-6.906645E-2,-2.3408173E-1,5.1842995E-2,-4.9602617E-2,-3.082546E-1],"split_indices":[147,154,22,121,132,380,479,0,482,321,205,75,136,0,156,0,0,379,0,4,0,46,94,0,0,70,273,0,0,433,159,0,0,144,451,0,0,128,494,0,0,112,490,293,0,0,186,75,291,44,306,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.2975085E2,1.0957213E2,2.0178726E1,6.3141212E0,1.0325801E2,8.876524E0,1.1302202E1,1.3162205E0,4.9979005E0,2.8364408E0,6.040083E0,5.491026E1,4.834775E1,1.2493852E0,1.0052817E1,3.8227143E0,1.1751863E0,4.1128855E0,1.9271972E0,8.168295E0,1.8845226E0,3.276403E1,1.5583722E1,1.8348104E0,1.0016304E0,5.082287E1,4.087396E0,2.9591367E0,1.1537488E0,5.3108034E0,2.7453228E1,1.5224838E0,2.5649123E0,1.4585066E1,3.62378E1,6.9597144E0,1.2085806E0,1.0813668E1,4.770054E0,4.1288853E0,1.1819181E0,7.591692E0,6.993374E0,2.4292226E0,2.3408313E0,1.8178797E0,8.995789E0,9.189132E0,1.8264095E1,3.1601425E1,4.6363754E0,2.4174075E0,5.1742845E0,1.0215106E0,1.407712E0,1.3619058E1,1.7982367E1,1.1316866E1,6.94723E0,1.9992914E0,2.637084E0,7.8237534E0,1.3653783E0,4.5768886E0,2.4164853E0,1.1893078E0,7.806481E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"67","size_leaf_vector":"0"}},{"base_weights":[-1.0724835E-2,2.0377645E-1,-1.4559579E-1,3.829644E-1,-1.5511353E-1,-3.3991674E-1,1.3336356E-1,5.864911E-1,-6.1391182E-2,2.3726024E-2,-5.5964446E-1,-4.3271053E-1,3.488151E-1,5.1736677E-1,-1.3374561E-1,-2.1094218E-1,7.045338E-1,-6.73305E-1,2.0712547E-1,7.2625643E-1,-2.1775474E-1,-8.009149E-2,6.181565E-1,-3.4297326E-1,4.9568573E-1,1.5797849E-1,-6.132248E-1,2.7524024E-1,-6.2677693E-1,1.3419162E-1,-6.415388E-1,1.0289912E-1,-1.8598607E-1,-5.574092E-2,-2.3402537E-1,2.4121644E-1,4.528077E-3,-7.092457E-1,1.16983674E-1,1.0858743E-2,-2.543249E-1,-2.0123675E-1,2.4112411E-1,9.874903E-2,-8.1979936E-1,2.384301E-1,-8.012535E-2,7.7755295E-2,-1.7106932E-1,-5.3156734E-1,1.7325528E-1,-3.700798E-2,9.225592E-1,8.5727096E-1,-9.539228E-2,-4.243156E-1,4.5005506E-1,5.36577E-2,-7.9190284E-1,-2.893501E-1,-3.3735346E-2,-3.602208E-2,9.4979495E-1,6.346377E-1,-1.2114697E-1,8.5108675E-2,-2.1554847E-1,3.0768907E-1,1.1053967E-1,-1.7301704E-1,1.5115865E-1,-1.487608E-1,1.3116397E-1,1.4113474E-1,-5.555228E-1,-8.35787E-3,-8.1287265E-1,-1.1517361E-2,-2.6747012E-1,2.4670523E-1,-3.4050774E-2,2.4585957E-2,-2.8481865E-1,-1.7825016E-1,1.4670734E-1,8.363782E-2,3.0134302E-1,-2.2086763E-1,-1.5724674E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,true,true,false,false,true,true,false,false,true,false,true,true,false,true,true,true,false,true,false,true,true,true,false,true,false,true,false,true,false,false,false,false,false,false,true,false,false,false,true,false,false,false,false,false,false,false,true,false,true,true,false,true,true,true,false,false,false,false,false,true,true,false,false,false,false,false,false,false,false,false,false,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":8,"left_children":[1,3,5,7,11,9,13,15,27,17,29,25,21,19,23,31,53,33,55,51,47,-1,35,49,45,-1,43,41,39,-1,37,-1,-1,-1,-1,-1,-1,57,-1,-1,-1,73,-1,-1,59,-1,-1,-1,-1,75,-1,71,67,61,69,65,63,-1,77,-1,-1,-1,85,79,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,87,83,81,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[3.493656E0,3.058745E0,4.0619645E0,2.8754282E0,2.4232166E0,3.5254593E0,3.2633786E0,2.1845002E0,2.221973E0,2.378543E0,2.5498028E0,2.21394E0,1.2137973E0,2.1514447E0,2.6933527E0,1.0188475E0,2.4116964E0,5.7231665E-2,2.4353867E0,1.5832405E0,7.599956E-1,0E0,5.148603E-1,2.906027E0,1.3960656E0,0E0,2.1489701E0,2.036385E0,6.6411436E-1,0E0,2.0775242E0,0E0,0E0,0E0,0E0,0E0,0E0,1.9440851E0,0E0,0E0,0E0,1.3832809E0,0E0,0E0,7.3122215E-1,0E0,0E0,0E0,0E0,1.8986752E0,0E0,8.8625693E-1,9.3967915E-2,1.5710678E0,1.4320216E0,1.1577382E0,1.9981546E0,0E0,1.6665449E0,0E0,0E0,0E0,2.6414394E-1,1.4241326E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,2.8370762E-1,1.9039425E0,1.1051078E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,13,13,12,12,14,14,11,11,8,8,10,10,15,15,17,17,22,22,30,30,28,28,27,27,26,26,24,24,20,20,23,23,19,19,16,16,18,18,37,37,44,44,53,53,56,56,55,55,52,52,54,54,51,51,41,41,49,49,58,58,63,63,76,76,75,75,62,62,74,74],"right_children":[2,4,6,8,12,10,14,16,28,18,30,26,22,20,24,32,54,34,56,52,48,-1,36,50,46,-1,44,42,40,-1,38,-1,-1,-1,-1,-1,-1,58,-1,-1,-1,74,-1,-1,60,-1,-1,-1,-1,76,-1,72,68,62,70,66,64,-1,78,-1,-1,-1,86,80,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,88,84,82,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[-2.4530811E0,4.1278556E-1,1.468729E0,3.0924356E-1,1.9619074E0,-3.5090437E0,-2.114107E0,-6.9320664E0,3.6453547E0,-7.299544E-1,-8.529522E0,-9.2961675E-1,-8.0017936E-1,3.634656E0,6.819783E-1,-5.464606E-1,6.7967625E0,-3.8622454E-1,-5.581887E-1,-9.11814E-1,2.1654013E-1,-8.009149E-2,1.0116103E-1,8.521781E0,3.9165115E-1,1.5797849E-1,-1.0358188E0,-8.56832E-1,-1.3261546E-1,1.3419162E-1,1.6601745E0,1.0289912E-1,-1.8598607E-1,-5.574092E-2,-2.3402537E-1,2.4121644E-1,4.528077E-3,-1.3666791E0,1.16983674E-1,1.0858743E-2,-2.543249E-1,-4.733117E-1,2.4112411E-1,9.874903E-2,4.0829053E0,2.384301E-1,-8.012535E-2,7.7755295E-2,-1.7106932E-1,-1.2694299E0,1.7325528E-1,3.271674E-1,6.0106546E-1,-1.3228234E0,7.37866E-2,-3.1363174E-1,1.0406269E0,5.36577E-2,-1.1926293E0,-2.893501E-1,-3.3735346E-2,-3.602208E-2,-9.634871E0,7.0381284E-1,-1.2114697E-1,8.5108675E-2,-2.1554847E-1,3.0768907E-1,1.1053967E-1,-1.7301704E-1,1.5115865E-1,-1.487608E-1,1.3116397E-1,1.4113474E-1,4.734554E0,-8.734884E-1,-1.4515672E0,-1.1517361E-2,-2.6747012E-1,2.4670523E-1,-3.4050774E-2,2.4585957E-2,-2.8481865E-1,-1.7825016E-1,1.4670734E-1,8.363782E-2,3.0134302E-1,-2.2086763E-1,-1.5724674E-2],"split_indices":[61,255,137,326,31,63,8,33,26,365,139,280,418,2,364,122,91,250,466,225,210,0,226,92,234,0,327,71,360,0,377,0,0,0,0,0,0,293,0,0,0,469,0,0,118,0,0,0,0,132,0,221,466,417,405,493,428,0,414,0,0,0,2,343,0,0,0,0,0,0,0,0,0,0,152,3,194,0,0,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.1873687E2,4.5639717E1,7.309715E1,3.039627E1,1.5243445E1,4.295589E1,3.0141262E1,2.0628157E1,9.768114E0,1.6515408E1,2.6440483E1,9.920118E0,5.323327E0,1.198065E1,1.8160612E1,2.5583453E0,1.806981E1,2.8838294E0,1.3631578E1,9.338977E0,2.6416728E0,1.6225473E0,3.70078E0,1.3971491E1,4.189121E0,1.2674304E0,8.652688E0,6.444332E0,3.3237822E0,1.653051E0,2.4787432E1,1.1552064E0,1.4031389E0,1.0104921E0,1.8733373E0,2.5740323E0,1.1267475E0,2.355376E1,1.2336713E0,1.0399306E0,2.2838516E0,3.7146304E0,2.7297018E0,1.4489911E0,7.203697E0,2.9790792E0,1.2100415E0,1.23511E0,1.4065628E0,1.19463825E1,2.0251086E0,2.0765212E0,7.2624555E0,1.5112547E1,2.957264E0,3.5798635E0,1.0051715E1,1.9064246E0,2.1647335E1,5.67851E0,1.5251871E0,1.2812082E0,1.3831339E1,8.47618E0,1.575535E0,1.0612808E0,2.5185828E0,5.4058657E0,1.8565897E0,1.6608443E0,1.2964196E0,1.0321512E0,1.0443699E0,1.1695044E0,2.545126E0,4.537199E0,7.4091835E0,2.6905587E0,1.8956776E1,6.6967344E0,1.7794455E0,1.0336883E0,6.3754954E0,1.9940495E0,2.5431495E0,1.5017502E0,1.2329589E1,1.5316439E0,1.0134821E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"89","size_leaf_vector":"0"}},{"base_weights":[-6.0583623E-3,4.4529295E-1,-6.5864146E-2,-2.2989488E-1,6.675768E-1,4.9330607E-2,-3.6292574E-1,-1.8222836E-1,8.7410524E-2,-8.142535E-2,4.1419896E-1,-3.227336E-2,7.819469E-1,5.9409756E-2,-6.17561E-1,-3.0864763E-1,2.4074526E-1,1.30994655E-2,9.037927E-1,4.195308E-3,-7.832685E-1,6.08451E-1,-1.4227296E-1,-2.3623581E-1,5.545235E-1,7.892075E-2,-4.7896704E-1,2.9067007E-1,7.7147506E-2,1.1384383E-1,-4.9563885E-1,2.2396411E-1,-1.0245565E-2,8.253044E-1,5.360822E-3,-5.5243933E-1,3.5728598E-1,1.4378545E-2,-8.7061805E-1,-2.4843004E-1,4.71444E-1,-4.043752E-1,1.5916923E-1,3.8314638E-1,-1.8814088E-1,-2.2455315E-1,2.447754E-2,1.8010558E-1,-2.7951008E-2,-1.796342E-1,-2.1316155E-3,6.0407585E-1,-9.148865E-2,1.9656356E-1,-2.1945067E-1,-1.571666E-2,-2.2739835E-1,-1.6179381E-2,-9.5832014E-1,3.959674E-1,-1.7433254E-1,-7.5889456E-1,-1.9385304E-1,2.7452683E-2,9.041625E-1,1.9826931E-1,-6.890919E-2,-1.0644314E-1,1.747913E-1,-1.3667946E-1,1.3805085E-1,-4.8752714E-2,-2.7174452E-1,3.0186436E-1,7.858459E-2,-3.7770145E-2,2.0394756E-1,-9.325484E-2,2.1227005E-1,-9.761361E-2,-3.1476948E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,true,true,true,false,false,true,true,false,false,false,true,true,true,false,true,true,false,false,true,true,true,true,true,false,false,false,true,false,false,false,false,true,true,false,true,true,false,true,false,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false,true,true,false,true,true,false,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":9,"left_children":[1,3,5,7,11,9,13,-1,-1,15,21,-1,17,23,19,25,39,-1,27,41,37,33,29,35,31,43,61,-1,-1,-1,45,-1,-1,63,59,55,47,-1,57,53,51,49,-1,65,-1,-1,-1,-1,-1,-1,-1,77,-1,67,-1,-1,-1,-1,79,75,-1,71,69,-1,73,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[2.9237976E0,2.0557327E0,3.2877545E0,9.4115275E-1,9.521904E-1,3.360653E0,2.9209344E0,0E0,0E0,3.8935628E0,2.0929973E0,0E0,7.389841E-1,1.7411927E0,1.7362051E0,2.0862896E0,2.615637E0,0E0,6.141424E-2,1.1896349E0,1.0063219E0,1.8197575E0,1.2230158E0,1.621591E0,5.241655E-1,2.459744E0,1.584804E0,0E0,0E0,0E0,6.128734E-1,0E0,0E0,5.105448E-1,1.3057349E0,5.095674E-1,4.112287E-1,0E0,8.131428E-1,1.8097663E0,1.768992E0,2.3671877E-1,0E0,1.4866465E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,1.4962983E0,0E0,1.2332362E0,0E0,0E0,0E0,0E0,2.8257275E-1,5.911951E-1,0E0,7.821598E-1,2.2799168E0,0E0,3.3579683E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,9,9,12,12,14,14,10,10,13,13,15,15,18,18,22,22,24,24,21,21,23,23,20,20,16,16,19,19,25,25,30,30,36,36,41,41,40,40,39,39,35,35,38,38,34,34,26,26,33,33,43,43,53,53,62,62,61,61,64,64,59,59,51,51,58,58],"right_children":[2,4,6,8,12,10,14,-1,-1,16,22,-1,18,24,20,26,40,-1,28,42,38,34,30,36,32,44,62,-1,-1,-1,46,-1,-1,64,60,56,48,-1,58,54,52,50,-1,66,-1,-1,-1,-1,-1,-1,-1,78,-1,68,-1,-1,-1,-1,80,76,-1,72,70,-1,74,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[-8.964586E0,-4.5199847E0,4.7324567E0,3.8571584E-1,-1.1280526E0,3.8640652E0,-1.0913877E0,-1.8222836E-1,8.7410524E-2,1.7236207E0,4.819636E-1,-3.227336E-2,-1.2158289E0,3.4370166E-1,-5.5305004E0,-5.0252855E-1,-4.0908303E0,1.30994655E-2,1.5253885E0,9.5978725E-1,-1.3690306E0,2.3884482E0,-1.680573E0,4.149622E0,6.1887974E-1,6.2840266E0,8.1120896E-1,2.9067007E-1,7.7147506E-2,1.1384383E-1,7.238129E-1,2.2396411E-1,-1.0245565E-2,-8.798786E0,5.1924458E0,-2.2539645E-2,3.09698E-1,1.4378545E-2,-1.3542114E0,9.095875E-2,9.223715E0,1.5816227E-2,1.5916923E-1,3.5816078E0,-1.8814088E-1,-2.2455315E-1,2.447754E-2,1.8010558E-1,-2.7951008E-2,-1.796342E-1,-2.1316155E-3,-1.4932576E0,-9.148865E-2,-5.0132793E-1,-2.1945067E-1,-1.571666E-2,-2.2739835E-1,-1.6179381E-2,-4.7064962E0,-2.96465E-1,-1.7433254E-1,-6.01981E0,4.4567582E-1,2.7452683E-2,1.06541E0,1.9826931E-1,-6.890919E-2,-1.0644314E-1,1.747913E-1,-1.3667946E-1,1.3805085E-1,-4.8752714E-2,-2.7174452E-1,3.0186436E-1,7.858459E-2,-3.7770145E-2,2.0394756E-1,-9.325484E-2,2.1227005E-1,-9.761361E-2,-3.1476948E-1],"split_indices":[64,129,134,54,208,156,86,0,0,78,287,0,378,251,4,269,73,0,178,104,185,27,25,131,211,147,36,0,0,0,459,0,0,79,141,497,260,0,421,189,23,187,0,132,0,0,0,0,0,0,0,328,0,429,0,0,0,0,116,273,0,164,293,0,458,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.06228676E2,1.1674692E1,9.4553986E1,2.883082E0,8.79161E0,6.8734146E1,2.5819838E1,1.5761971E0,1.3068848E0,5.1172546E1,1.75616E1,1.1380113E0,7.6535983E0,1.0000234E1,1.5819604E1,3.0032814E1,2.1139734E1,1.276635E0,6.376963E0,3.5340533E0,1.228555E1,1.2948246E1,4.613354E0,6.58877E0,3.4114637E0,9.332646E0,2.0700167E1,5.3408384E0,1.0361248E0,1.8329961E0,2.780358E0,2.3701499E0,1.0413139E0,9.252766E0,3.6954808E0,4.343796E0,2.244974E0,1.2112172E0,1.1074333E1,6.75004E0,1.4389694E1,2.1201472E0,1.4139062E0,6.841938E0,2.4907086E0,1.7270498E0,1.0533082E0,1.2420169E0,1.002957E0,1.0829585E0,1.0371888E0,1.2479762E1,1.9099318E0,3.822213E0,2.9278271E0,1.6309872E0,2.7128086E0,1.2305281E0,9.843805E0,2.4131062E0,1.2823745E0,9.606597E0,1.109357E1,1.1075107E0,8.145255E0,4.6541743E0,2.1877637E0,1.6074715E0,2.2147415E0,8.139476E0,2.9540937E0,2.31511E0,7.2914867E0,6.539323E0,1.6059324E0,1.0461828E0,1.3669235E0,1.0645783E0,1.1415184E1,1.8114238E0,8.032381E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"81","size_leaf_vector":"0"}},{"base_weights":[-7.6359296E-3,-1.21141255E-1,2.1417494E-1,-2.5172455E-2,-4.905114E-1,-1.1978601E-2,5.2540463E-1,1.2253718E-1,-3.2409528E-1,2.0021336E-1,-5.658567E-1,1.5474895E-1,-6.4252925E-1,-1.4935338E-1,7.139636E-1,3.6941966E-1,-2.4087997E-1,-2.723724E-1,4.3301785E-1,1.0679977E-1,-1.9396804E-1,3.6483298E-3,-7.6843727E-1,2.7431586E-1,1.8034852E-1,-9.348529E-2,1.473568E-1,2.2128856E-1,-5.146711E-1,2.9436229E-2,-7.5684524E-1,-1.02899574E-1,5.61853E-1,1.2736627E-1,-5.507E-1,-1.2222522E-1,5.652555E-1,-6.761701E-2,-2.5217736E-1,-1.0370397E-1,1.5651034E-1,-1.369615E-1,-8.6299086E-1,-2.6304054E-1,5.74706E-3,9.799629E-2,-6.407144E-1,-1.3920319E-1,6.0024154E-1,3.3312425E-1,-6.121526E-1,-7.3197163E-3,-2.2570513E-1,-4.249008E-1,3.8398746E-1,7.580369E-1,-1.012781E-1,-2.8695056E-1,-6.1392654E-2,3.8259894E-2,-2.4726973E-1,-5.299211E-1,5.192713E-1,-1.08629964E-1,6.617481E-1,2.222239E-1,1.8960565E-2,-5.19061E-2,1.7139606E-1,-2.1819675E-1,3.3047095E-2,2.0016839E-1,2.7039118E-2,-2.3508237E-1,1.7121546E-2,2.2190456E-1,-4.376479E-2,-1.0039929E-2,2.6803696E-1,-2.1436967E-1,-5.353361E-2,-7.240748E-2,1.8801208E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,true,true,true,false,true,true,true,true,true,false,false,false,true,false,false,false,false,false,true,false,true,false,false,true,true,false,true,false,false,false,true,false,false,false,false,false,false,false,true,false,false,true,true,false,true,true,false,false,false,false,true,true,false,false,false,false,false,false,true,false,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":10,"left_children":[1,3,5,7,11,9,13,15,27,17,29,25,21,19,23,31,45,33,47,-1,-1,-1,43,-1,39,-1,-1,35,41,-1,37,49,63,-1,51,-1,65,-1,-1,-1,-1,61,57,-1,-1,53,59,-1,55,67,79,-1,-1,69,81,77,-1,-1,-1,-1,-1,73,71,-1,75,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[2.471037E0,2.2875729E0,2.3456092E0,2.3602073E0,1.4066603E0,2.4904706E0,1.8890555E0,3.309946E0,1.9489195E0,1.8081356E0,8.178649E-1,6.782097E-1,9.1441584E-1,1.2067894E0,1.0246234E0,2.036666E0,2.1125004E0,1.3363663E0,1.8803418E0,0E0,0E0,0E0,8.0407E-1,0E0,9.493279E-1,0E0,0E0,1.3701948E0,1.6426644E0,0E0,1.8329144E-2,1.7910606E0,1.681705E0,0E0,4.5892918E-1,0E0,2.9027152E-1,0E0,0E0,0E0,0E0,2.237534E0,2.2130394E-1,0E0,0E0,1.5173272E0,1.1022239E0,0E0,1.6523952E0,6.411773E-1,3.4586787E-2,0E0,0E0,6.7499655E-1,1.1194904E0,9.287181E-1,0E0,0E0,0E0,0E0,0E0,8.2719445E-1,1.5289909E-1,0E0,1.0176291E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,13,13,12,12,14,14,11,11,8,8,10,10,15,15,17,17,27,27,30,30,24,24,28,28,22,22,16,16,18,18,31,31,34,34,45,45,48,48,42,42,46,46,41,41,32,32,36,36,49,49,53,53,62,62,61,61,64,64,55,55,50,50,54,54],"right_children":[2,4,6,8,12,10,14,16,28,18,30,26,22,20,24,32,46,34,48,-1,-1,-1,44,-1,40,-1,-1,36,42,-1,38,50,64,-1,52,-1,66,-1,-1,-1,-1,62,58,-1,-1,54,60,-1,56,68,80,-1,-1,70,82,78,-1,-1,-1,-1,-1,74,72,-1,76,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[4.0571138E-1,6.2840266E0,1.7952664E0,3.948227E0,-7.502468E0,4.5300994E0,-1.1436375E0,3.3679242E0,-5.073654E-1,-3.233758E-1,-7.049883E0,-1.7046089E0,-6.4081664E0,4.731121E-1,5.9370106E-1,-4.828108E0,1.5662811E0,-8.1345725E-1,-1.1188068E0,1.0679977E-1,-1.9396804E-1,3.6483298E-3,9.2623323E-1,2.7431586E-1,8.1120896E-1,-9.348529E-2,1.473568E-1,-4.6069667E-1,1.3805943E0,2.9436229E-2,-6.4803445E-1,1.4104366E0,-1.3970176E0,1.2736627E-1,-3.2028374E-1,-1.2222522E-1,-2.1802232E-1,-6.761701E-2,-2.5217736E-1,-1.0370397E-1,1.5651034E-1,4.058962E-1,5.119229E0,-2.6304054E-1,5.74706E-3,-5.6050596E0,-9.505754E-1,-1.3920319E-1,1.2707098E0,-3.9600158E-1,-2.1579285E0,-7.3197163E-3,-2.2570513E-1,-2.6548202E0,-9.098961E-1,-5.2664247E0,-1.012781E-1,-2.8695056E-1,-6.1392654E-2,3.8259894E-2,-2.4726973E-1,1.847589E0,1.6286445E-1,-1.08629964E-1,1.514138E0,2.222239E-1,1.8960565E-2,-5.19061E-2,1.7139606E-1,-2.1819675E-1,3.3047095E-2,2.0016839E-1,2.7039118E-2,-2.3508237E-1,1.7121546E-2,2.2190456E-1,-4.376479E-2,-1.0039929E-2,2.6803696E-1,-2.1436967E-1,-5.353361E-2,-7.240748E-2,1.8801208E-1],"split_indices":[264,147,133,26,159,140,415,110,212,197,47,12,81,168,284,99,61,285,307,0,0,0,379,0,36,0,0,294,117,0,209,123,471,0,386,0,369,0,0,0,0,161,16,0,0,71,173,0,177,252,151,0,0,152,259,99,0,0,0,0,0,3,152,0,252,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[9.611717E1,6.3881504E1,3.223567E1,5.1554813E1,1.2326691E1,1.9226406E1,1.3009262E1,3.4899605E1,1.6655209E1,1.436259E1,4.863816E0,2.3469179E0,9.979773E0,2.8867955E0,1.0122467E1,2.0771948E1,1.4127657E1,4.684085E0,9.678505E0,1.5725172E0,1.3142784E0,1.7554222E0,8.22435E0,6.8398805E0,3.2825868E0,1.0144663E0,1.3324516E0,4.271113E0,1.2384095E1,1.1951362E0,3.6686792E0,6.1477733E0,1.4624175E1,1.1868532E0,3.4972315E0,1.4451635E0,2.8259494E0,1.0019051E0,2.6667743E0,1.2868536E0,1.9957333E0,6.6097975E0,5.774298E0,7.1300635E0,1.0942864E0,8.055335E0,6.072322E0,1.2418945E0,8.43661E0,3.498341E0,2.6494322E0,1.2792994E0,2.2179322E0,2.6763217E0,5.3790135E0,7.3856544E0,1.050956E0,4.6595783E0,1.1147199E0,1.2296784E0,4.8426437E0,4.2545843E0,2.355213E0,1.1707588E0,1.3453416E1,1.7417097E0,1.0842397E0,1.2070636E0,2.2912776E0,1.4803015E0,1.1960202E0,1.3388226E0,1.0163903E0,2.7368283E0,1.517756E0,1.2346183E1,1.1072328E0,1.1890042E0,6.19665E0,1.6194494E0,1.0299828E0,1.5055926E0,3.873421E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"83","size_leaf_vector":"0"}},{"base_weights":[-1.7076256E-2,-1.06007926E-1,2.5770286E-1,-2.0001848E-1,3.3590546E-1,-2.9133338E-1,4.7314337E-1,3.9463657E-1,-2.554535E-1,-4.880555E-1,1.0786799E-1,-1.1820677E-1,4.7431946E-1,-1.5638098E-1,6.07143E-1,-6.260715E-2,6.0378957E-1,-2.1410413E-1,3.787004E-2,7.2291046E-1,-1.7342351E-2,-1.5422133E-1,8.1621535E-2,3.74516E-1,-3.1257024E-1,-1.0884206E-1,6.449399E-1,3.507457E-2,2.2315957E-1,-6.6678226E-2,1.8442777E-1,6.763387E-2,2.5575978E-1,8.592683E-2,-1.4331698E-1,3.3935197E-2,8.3017033E-1,-4.5260233E-1,-3.131267E-2,1.3814752E-1,-9.56122E-2,-2.3811623E-2,-5.913424E-1,7.507208E-2,2.7715826E-1,2.1692084E-1,-3.7748516E-1,-1.0724852E-1,1.3915634E-1,-5.042662E-2,1.558992E-1,7.5454965E-2,-1.9956334E-1,-1.8580812E-1,8.331676E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false,true,false,false,false,true,false,false,false,false,true,false,false,false,false,false,false,false,false,true,true,false,false,false,false,false,false,true,false,false,false,false,false,false,true,false,false,true,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":11,"left_children":[1,3,5,7,11,9,13,15,23,17,-1,-1,25,21,19,-1,27,-1,-1,35,-1,-1,-1,29,37,33,31,-1,-1,-1,-1,39,-1,-1,-1,-1,43,41,45,-1,-1,47,51,-1,-1,49,53,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[2.157814E0,2.8330367E0,2.693877E0,1.9311376E0,1.4045081E0,1.0345001E0,1.4362476E0,7.7304786E-1,1.9663904E0,8.6290085E-1,0E0,0E0,1.1233432E0,6.8939143E-1,1.0667553E0,0E0,1.9022644E-1,0E0,0E0,6.526699E-1,0E0,0E0,0E0,8.1441605E-1,1.8632655E0,6.109074E-1,9.4100857E-1,0E0,0E0,0E0,0E0,6.46127E-1,0E0,0E0,0E0,0E0,2.8766203E-1,1.8747325E0,1.5357527E0,0E0,0E0,1.5984026E0,1.658225E0,0E0,0E0,1.309862E0,1.3495916E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,14,14,13,13,8,8,12,12,16,16,23,23,26,26,25,25,19,19,24,24,31,31,37,37,36,36,38,38,41,41,45,45,42,42,46,46],"right_children":[2,4,6,8,12,10,14,16,24,18,-1,-1,26,22,20,-1,28,-1,-1,36,-1,-1,-1,30,38,34,32,-1,-1,-1,-1,40,-1,-1,-1,-1,44,42,46,-1,-1,48,52,-1,-1,50,54,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[5.923786E0,6.9030175E0,-5.0228133E0,-1.0723837E1,-8.186522E0,3.1108597E-1,-8.719484E0,-2.7605066E0,-1.0502298E1,2.3606516E-1,1.0786799E-1,-1.1820677E-1,-6.386464E0,8.9657784E-1,9.498953E-1,-6.260715E-2,-1.742125E0,-2.1410413E-1,3.787004E-2,-9.967978E-1,-1.7342351E-2,-1.5422133E-1,8.1621535E-2,-3.7384152E0,3.8893974E-1,-2.0504887E-1,-5.181928E0,3.507457E-2,2.2315957E-1,-6.6678226E-2,1.8442777E-1,-5.8213852E-2,2.5575978E-1,8.592683E-2,-1.4331698E-1,3.3935197E-2,-6.389453E0,-5.3744955E0,2.646595E-1,1.3814752E-1,-9.56122E-2,1.483927E-1,-1.129789E1,7.507208E-2,2.7715826E-1,-1.7625952E0,4.075739E0,-1.0724852E-1,1.3915634E-1,-5.042662E-2,1.558992E-1,7.5454965E-2,-1.9956334E-1,-1.8580812E-1,8.331676E-2],"split_indices":[161,3,5,57,16,400,13,1,121,322,0,0,6,38,276,0,74,0,0,186,0,0,0,31,264,429,123,0,0,0,0,411,0,0,0,0,35,81,180,0,0,357,147,0,0,16,53,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[8.618474E1,6.5575554E1,2.060918E1,5.45238E1,1.1051753E1,5.7086887E0,1.4900492E1,4.127622E0,5.039618E1,4.575878E0,1.1328104E0,1.4672889E0,9.584464E0,2.5857449E0,1.2314747E1,1.0631242E0,3.064498E0,3.2283313E0,1.3475469E0,1.0414436E1,1.9003109E0,1.3003379E0,1.285407E0,3.7274308E0,4.6668747E1,2.2514255E0,7.3330383E0,1.0945885E0,1.9699094E0,1.0913389E0,2.636092E0,2.2901187E0,5.0429196E0,1.1947546E0,1.0566709E0,1.8651732E0,8.549263E0,3.0749851E1,1.5918898E1,1.0403204E0,1.2497983E0,7.803609E0,2.2946241E1,1.6996444E0,6.8496184E0,9.488357E0,6.430541E0,4.79178E0,3.0118287E0,4.372486E0,5.1158705E0,1.6551925E0,2.1291048E1,4.7366714E0,1.6938697E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"55","size_leaf_vector":"0"}},{"base_weights":[-4.6404456E-3,1.1793071E-1,-2.0409165E-1,1.9603664E-1,-3.2996526E-1,3.399558E-1,-2.959365E-1,5.192491E-1,4.9721025E-2,5.986675E-1,-1.011317E-1,1.272128E-1,-5.555674E-1,-4.541876E-1,1.6596353E-1,-7.1822226E-2,6.28671E-1,4.7735892E-2,2.190032E-1,-1.5890247E-1,-6.503221E-1,-3.046176E-1,4.6553722E-1,-2.3162378E-1,2.1090744E-2,-2.4291088E-1,2.5837827E-1,-1.1560946E-1,1.6079876E-1,7.1592534E-1,-4.3604262E-2,-7.7707726E-1,-1.08748116E-1,1.8994954E-1,-4.5415992E-1,-4.0996578E-1,4.120202E-1,2.0224212E-1,-3.452471E-2,-5.105589E-1,2.1751241E-1,-1.8744437E-1,7.50114E-2,8.1637704E-1,2.3353187E-2,8.434783E-2,-1.5074256E-1,5.860481E-1,-1.5433644E-1,-1.0102224E-1,4.9799308E-1,7.926673E-3,-1.8944359E-1,4.445511E-2,-2.1770783E-1,-6.0825163E-1,4.496437E-2,-3.2723125E-2,-2.56816E-1,-2.8227073E-1,1.7964792E-1,1.0787036E-1,2.7392632E-1,1.3669853E-1,-1.7399013E-1,1.8066695E-2,-1.4742443E-1,-2.3123755E-1,1.8459508E-2,-5.580797E-2,2.0654388E-1,2.0970963E-1,-9.6728355E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,true,true,true,true,true,false,false,true,false,false,true,false,true,false,false,true,true,true,true,false,false,true,true,false,false,true,false,false,true,true,true,false,true,false,false,true,true,false,false,true,false,false,false,true,true,false,true,false,false,false,false,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":12,"left_children":[1,3,5,7,11,9,13,15,25,17,-1,27,23,19,21,-1,29,-1,-1,39,31,41,37,-1,-1,33,35,-1,-1,43,-1,57,45,59,55,51,47,-1,-1,53,49,-1,-1,61,-1,-1,-1,69,63,-1,71,-1,-1,-1,-1,67,-1,-1,-1,65,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[1.9924576E0,1.831465E0,1.6502728E0,2.037562E0,8.825901E-1,1.0679587E0,2.05398E0,1.2704191E0,1.9642154E0,1.1118448E-1,0E0,9.2204785E-1,7.937796E-1,1.079803E0,1.2188152E0,0E0,9.189496E-1,0E0,0E0,1.3654877E0,7.3595905E-1,7.9135203E-1,6.908779E-1,0E0,0E0,1.3186712E0,2.0853686E0,0E0,0E0,6.423006E-1,0E0,4.0898705E-1,6.6776407E-1,1.1586303E0,9.6164703E-1,4.0235275E-1,1.6352675E0,0E0,0E0,8.4422684E-1,9.8828006E-1,0E0,0E0,8.736038E-2,0E0,0E0,0E0,1.0927477E0,1.4262955E0,0E0,4.3915474E-1,0E0,0E0,0E0,0E0,8.801546E-1,0E0,0E0,0E0,2.4864478E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,13,13,14,14,12,12,8,8,11,11,16,16,20,20,25,25,26,26,22,22,19,19,21,21,29,29,32,32,36,36,40,40,35,35,39,39,34,34,31,31,33,33,43,43,48,48,59,59,55,55,47,47,50,50],"right_children":[2,4,6,8,12,10,14,16,26,18,-1,28,24,20,22,-1,30,-1,-1,40,32,42,38,-1,-1,34,36,-1,-1,44,-1,58,46,60,56,52,48,-1,-1,54,50,-1,-1,62,-1,-1,-1,70,64,-1,72,-1,-1,-1,-1,68,-1,-1,-1,66,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[2.0757966E0,7.6879067E0,-1.1460266E0,-3.6159916E0,-5.5506575E-1,5.477701E-1,5.9728175E-1,-8.626573E0,-2.4439175E0,6.021564E-1,-1.011317E-1,6.7710745E-1,3.7718365E0,-6.564326E-2,4.792705E-2,-7.1822226E-2,1.0942316E0,4.7735892E-2,2.190032E-1,-1.9693625E0,7.535811E-1,3.1442222E-1,1.2500474E0,-2.3162378E-1,2.1090744E-2,-3.5562167E0,-7.625686E0,-1.1560946E-1,1.6079876E-1,6.225357E0,-4.3604262E-2,-1.0962812E0,3.0176878E-1,5.7507205E-1,8.538685E-1,-2.9847949E0,5.272397E0,2.0224212E-1,-3.452471E-2,-6.3230467E-1,-4.960393E-1,-1.8744437E-1,7.50114E-2,-5.4545844E-1,2.3353187E-2,8.434783E-2,-1.5074256E-1,-1.3674928E0,-9.121523E-1,-1.0102224E-1,3.866547E0,7.926673E-3,-1.8944359E-1,4.445511E-2,-2.1770783E-1,7.709378E-1,4.496437E-2,-3.2723125E-2,-2.56816E-1,-3.3047965E-1,1.7964792E-1,1.0787036E-1,2.7392632E-1,1.3669853E-1,-1.7399013E-1,1.8066695E-2,-1.4742443E-1,-2.3123755E-1,1.8459508E-2,-5.580797E-2,2.0654388E-1,2.0970963E-1,-9.6728355E-3],"split_indices":[117,75,260,159,253,186,391,137,151,156,0,17,36,381,252,0,314,0,0,10,438,264,391,0,0,144,156,0,0,106,0,411,468,47,364,23,9,0,0,315,204,0,0,323,0,0,0,338,74,0,136,0,0,0,0,381,0,0,0,457,0,0,0,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[7.951669E1,4.950346E1,3.0013233E1,4.263152E1,6.8719425E0,3.9447703E0,2.6068462E1,1.2490677E1,3.0140842E1,2.9379404E0,1.00683E0,2.414707E0,4.4572353E0,1.9428709E1,6.6397524E0,1.4247197E0,1.1065957E1,1.1847112E0,1.7532291E0,8.476822E0,1.0951887E1,2.576172E0,4.0635805E0,3.1380548E0,1.3191806E0,1.2477531E1,1.766331E1,1.0967577E0,1.3179493E0,1.0012331E1,1.0536258E0,8.522379E0,2.4295084E0,4.1253333E0,8.352199E0,2.989712E0,1.4673597E1,2.8710191E0,1.1925614E0,4.1982946E0,4.2785273E0,1.5530952E0,1.0230769E0,8.409083E0,1.6032479E0,1.3643231E0,1.0651853E0,1.1197962E1,3.475635E0,1.3702377E0,2.9082894E0,1.3031601E0,1.6865519E0,1.1101631E0,3.0881314E0,6.649047E0,1.7031515E0,1.1528038E0,7.369575E0,2.05807E0,2.0672634E0,2.2935514E0,6.115532E0,1.3994452E0,2.07619E0,1.0502461E0,1.0078238E0,5.229232E0,1.4198152E0,1.2160327E0,9.981929E0,1.8779303E0,1.0303591E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"73","size_leaf_vector":"0"}},{"base_weights":[-7.721405E-3,4.927694E-2,-4.1027224E-1,-5.5644186E-3,4.9803254E-1,2.3847035E-1,-6.3148504E-1,1.5916467E-1,-1.732136E-1,-2.8010415E-2,1.347771E-1,6.621731E-1,-7.43302E-2,3.1754177E-2,-7.814493E-1,-3.449906E-2,5.0840306E-1,2.477941E-1,2.136685E-2,2.8621438E-1,-3.1956914E-1,-2.5629595E-1,-6.864452E-2,1.6130291E-1,-3.8878772E-1,5.030745E-1,-1.0300194E-1,6.1498475E-1,-3.0328473E-2,-5.742407E-1,-9.4479986E-2,3.4950835E-1,-2.8889194E-1,7.778862E-3,2.1484323E-1,-2.671703E-1,4.3333516E-1,-6.791365E-1,5.7131156E-2,-2.1203299E-1,9.020918E-2,2.3488777E-2,7.3778915E-1,1.6173925E-1,-3.965667E-2,5.3824786E-2,-1.3403866E-1,-1.0342972E-1,1.718051E-1,5.224165E-2,2.4221507E-1,-2.3378497E-1,-9.186815E-3,-1.7985794E-1,2.5468696E-2,1.7645702E-1,4.0024216E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,true,false,true,true,true,false,false,false,false,false,true,false,true,false,false,true,false,false,false,false,false,false,false,true,false,true,false,true,true,false,false,true,true,true,false,false,true,false,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":13,"left_children":[1,3,5,7,11,9,13,15,19,-1,-1,17,-1,-1,21,23,27,-1,-1,25,29,-1,-1,31,39,33,-1,41,-1,37,35,43,53,-1,-1,45,55,51,-1,-1,47,-1,49,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[1.6959387E0,1.5991815E0,1.4659445E0,1.6512194E0,1.0301656E0,2.557404E-1,8.687887E-1,2.0584571E0,2.082892E0,0E0,0E0,5.288665E-1,0E0,0E0,4.7223568E-2,1.4742483E0,7.6555705E-1,0E0,0E0,1.2216289E0,1.2604048E0,0E0,0E0,1.2712382E0,1.2115978E0,6.036278E-1,0E0,5.58954E-1,0E0,9.766047E-1,1.334517E0,1.0022609E0,5.8725315E-1,0E0,0E0,9.379599E-1,2.3133904E-1,5.936327E-1,0E0,0E0,1.0145624E0,0E0,1.5345526E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,11,11,8,8,14,14,15,15,19,19,16,16,20,20,23,23,25,25,30,30,29,29,24,24,27,27,31,31,35,35,40,40,42,42,37,37,32,32,36,36],"right_children":[2,4,6,8,12,10,14,16,20,-1,-1,18,-1,-1,22,24,28,-1,-1,26,30,-1,-1,32,40,34,-1,42,-1,38,36,44,54,-1,-1,46,56,52,-1,-1,48,-1,50,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[7.8408446E0,9.250621E0,-5.6451964E0,-2.6823783E-1,6.837756E-1,-1.6865698E-1,-9.177927E0,2.9781587E0,-5.6589746E0,-2.8010415E-2,1.347771E-1,7.1997136E-1,-7.43302E-2,3.1754177E-2,5.301328E0,5.305309E-1,1.0024862E1,2.477941E-1,2.136685E-2,5.9645624E0,-2.792343E-1,-2.5629595E-1,-6.864452E-2,2.2402613E0,1.5300907E-1,-1.8682592E-1,-1.0300194E-1,-6.097951E0,-3.0328473E-2,8.444216E0,5.592025E0,5.857833E-1,8.1756485E-1,7.778862E-3,2.1484323E-1,-4.6005163E0,1.08087756E-1,7.5252905E0,5.7131156E-2,-2.1203299E-1,9.742692E-2,2.3488777E-2,-5.0876737E0,1.6173925E-1,-3.965667E-2,5.3824786E-2,-1.3403866E-1,-1.0342972E-1,1.718051E-1,5.224165E-2,2.4221507E-1,-2.3378497E-1,-9.186815E-3,-1.7985794E-1,2.5468696E-2,1.7645702E-1,4.0024216E-3],"split_indices":[27,45,6,147,367,323,61,16,57,0,0,323,0,0,46,337,40,0,0,109,12,0,0,98,284,332,0,110,0,39,161,345,153,0,0,111,219,149,0,0,262,0,154,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[7.203297E1,6.3867386E1,8.165581E0,5.7792324E1,6.0750623E0,2.0565116E0,6.109069E0,2.9175556E1,2.8616768E1,1.0152569E0,1.0412548E0,5.07111E0,1.0039526E0,1.0822299E0,5.026839E0,1.9347803E1,9.827752E0,3.659114E0,1.4119959E0,6.682856E0,2.1933912E1,3.9589121E0,1.0679268E0,1.2811876E1,6.535927E0,5.1200686E0,1.5627874E0,8.357322E0,1.4704307E0,9.562965E0,1.2370947E1,9.192552E0,3.619325E0,1.926835E0,3.1932335E0,9.693312E0,2.6776354E0,8.507766E0,1.0552E0,3.6423595E0,2.8935673E0,1.8611503E0,6.496171E0,6.5104494E0,2.6821022E0,2.7808053E0,6.912506E0,1.6717882E0,1.2217791E0,1.0988333E0,5.397338E0,7.195029E0,1.3127369E0,1.6468198E0,1.9725051E0,1.6637317E0,1.0139037E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"57","size_leaf_vector":"0"}},{"base_weights":[-3.776916E-3,-9.2178024E-2,2.6223204E-1,5.573676E-2,-3.069424E-1,-2.4241574E-1,4.3199736E-1,-1.5697564E-1,2.564173E-1,-1.5245147E-1,4.178197E-2,2.3341775E-1,-4.3203726E-1,-1.18154466E-1,5.530277E-1,-3.67662E-1,1.7842118E-1,-7.203384E-2,4.5157322E-1,6.444555E-1,-3.609035E-2,-1.3083838E-1,7.0336185E-2,-5.0188705E-2,5.0210977E-1,9.057129E-2,-5.300493E-1,9.278274E-2,-4.8719743E-1,2.245087E-2,1.8559644E-1,1.4817004E-1,-9.066298E-2,6.358956E-1,-6.407804E-2,-6.356869E-1,1.7314827E-2,1.898893E-1,-1.6319801E-1,1.6268711E-1,-8.111846E-2,7.3854405E-1,3.3287343E-2,-5.579916E-2,-2.1369149E-1,-7.5204694E-1,-1.01115465E-1,1.1825241E-1,-3.4205368E-1,5.3868577E-2,2.4578202E-1,-1.853096E-1,4.190283E-1,-4.772395E-2,2.3327586E-1,1.1086807E-1,-1.1761726E-1,-1.00610584E-1,1.13256015E-1,8.130035E-2,-1.476196E-1,2.4342497E-3,1.8683764E-1,2.3076972E-2,-1.07646264E-1,-6.10463E-2,-2.5459102E-1,-1.6934475E-1,3.6697168E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,true,false,true,true,true,false,false,false,true,false,true,false,true,false,false,true,true,false,false,false,true,true,false,true,false,false,false,false,false,false,true,false,true,false,false,false,false,true,true,false,true,false,false,true,false,false,false,false,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":14,"left_children":[1,3,5,7,11,9,13,15,23,-1,-1,17,25,21,19,27,39,-1,29,41,-1,-1,-1,37,33,31,35,-1,43,-1,-1,-1,-1,53,-1,45,55,51,-1,-1,47,49,-1,57,-1,65,59,-1,67,-1,-1,63,61,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[1.6067514E0,1.6452451E0,1.5651431E0,1.3719304E0,1.5241897E0,5.710169E-1,9.3861246E-1,1.1795659E0,1.263021E0,0E0,0E0,5.844015E-1,9.417801E-1,4.597951E-1,7.501552E-1,9.605869E-1,6.819581E-1,0E0,1.7410511E-1,4.088316E-1,0E0,0E0,0E0,1.0776613E0,1.0179684E0,7.211127E-1,8.690486E-1,0E0,8.2407045E-1,0E0,0E0,0E0,0E0,1.0263329E0,0E0,6.993284E-1,6.314286E-1,6.168247E-1,0E0,0E0,7.317208E-1,2.1997452E-1,0E0,6.275371E-1,0E0,3.6148977E-1,6.255951E-1,0E0,4.4528314E-1,0E0,0E0,1.6734403E-1,3.4662914E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,11,11,14,14,13,13,8,8,12,12,15,15,18,18,25,25,24,24,26,26,23,23,16,16,19,19,28,28,35,35,40,40,41,41,37,37,33,33,36,36,43,43,46,46,52,52,51,51,45,45,48,48],"right_children":[2,4,6,8,12,10,14,16,24,-1,-1,18,26,22,20,28,40,-1,30,42,-1,-1,-1,38,34,32,36,-1,44,-1,-1,-1,-1,54,-1,46,56,52,-1,-1,48,50,-1,58,-1,66,60,-1,68,-1,-1,64,62,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[5.0261116E0,1.1149788E0,-5.318883E0,-7.105297E-1,-7.874101E0,-1.117568E0,-7.0280857E0,1.6106155E0,-4.403355E-2,-1.5245147E-1,4.178197E-2,-6.3197274E0,-8.021965E0,-2.0899153E-1,1.0458261E0,-9.128267E0,-1.0367217E0,-7.203384E-2,-6.8589575E-3,6.5223885E0,-3.609035E-2,-1.3083838E-1,7.0336185E-2,3.5037208E0,1.2497944E0,-1.5732079E0,7.3944697E0,9.278274E-2,-1.4496181E0,2.245087E-2,1.8559644E-1,1.4817004E-1,-9.066298E-2,-1.0217129E0,-6.407804E-2,8.126664E-1,1.721282E-1,-5.1616997E-1,-1.6319801E-1,1.6268711E-1,-1.0648396E1,-6.9844227E0,3.3287343E-2,5.499625E-1,-2.1369149E-1,-8.3138216E-1,3.380486E-2,1.1825241E-1,2.4323592E-1,5.3868577E-2,2.4578202E-1,-7.4479795E-1,-4.6046847E-1,-4.772395E-2,2.3327586E-1,1.1086807E-1,-1.1761726E-1,-1.00610584E-1,1.13256015E-1,8.130035E-2,-1.476196E-1,2.4342497E-3,1.8683764E-1,2.3076972E-2,-1.07646264E-1,-6.10463E-2,-2.5459102E-1,-1.6934475E-1,3.6697168E-2],"split_indices":[5,81,160,161,61,108,73,47,391,0,0,78,51,305,498,9,60,0,473,15,0,0,0,109,468,20,36,0,86,0,0,0,0,469,0,374,430,291,0,0,167,76,0,414,0,204,330,0,169,0,0,14,293,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[6.6299446E1,5.0252743E1,1.6046703E1,3.0196184E1,2.0056559E1,3.931601E0,1.2115103E1,1.4764383E1,1.5431802E1,2.1224453E0,1.8091557E0,3.6075299E0,1.6449028E1,2.1889298E0,9.926173E0,9.008997E0,5.755386E0,1.1060584E0,2.5014715E0,8.777239E0,1.1489341E0,1.0264453E0,1.1624845E0,7.2188883E0,8.2129135E0,2.6097162E0,1.3839312E1,1.1138526E0,7.8951445E0,1.2133965E0,1.288075E0,1.1605538E0,1.4491624E0,7.0134115E0,1.199502E0,1.1465239E1,2.374073E0,5.271464E0,1.9474243E0,1.9425457E0,3.8128405E0,7.134621E0,1.6426176E0,3.1350265E0,4.760118E0,9.081622E0,2.383616E0,1.1703771E0,2.6424632E0,1.307252E0,5.827369E0,2.0712376E0,3.2002263E0,1.0423619E0,5.97105E0,1.3283461E0,1.0457269E0,2.0429807E0,1.0920458E0,1.378416E0,1.0052E0,1.4055625E0,1.7946638E0,1.0464194E0,1.0248182E0,1.8248019E0,7.25682E0,1.6409221E0,1.0015411E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"69","size_leaf_vector":"0"}},{"base_weights":[-2.423463E-3,-1.3054945E-1,1.8243015E-1,-2.790983E-1,1.4216994E-1,-6.2804736E-2,3.8325378E-1,-4.7098008E-1,6.545646E-2,2.8988823E-1,-2.8463137E-1,4.886494E-1,-7.7471964E-2,5.398752E-1,-9.203959E-3,9.987462E-2,-5.840567E-1,-6.811813E-2,1.8228261E-1,6.3044125E-1,-1.8645624E-2,-3.0810717E-1,2.4962695E-1,-1.11821406E-1,3.1626204E-1,-2.2008458E-4,1.9522381E-1,-3.1277984E-1,3.0854294E-1,-4.572212E-1,6.623818E-2,-6.925992E-1,4.3117963E-2,-4.934946E-1,5.9260186E-2,-7.117387E-2,4.93678E-1,3.2858398E-2,-1.6346568E-1,4.647232E-2,-6.054664E-1,1.5265906E-1,-7.950072E-3,1.668356E-1,7.720793E-1,-6.969227E-2,1.5293838E-1,-3.191635E-2,-7.666974E-1,2.503448E-3,-1.927057E-1,-4.8288498E-2,1.270997E-1,6.176983E-2,2.5125375E-1,-2.1010314E-1,-3.304957E-2,-8.854166E-2,1.0878457E-1,-2.0140395E-2,2.0131119E-1,-2.4775034E-1,-6.729467E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,true,false,false,false,true,true,true,true,true,false,true,true,false,false,false,false,true,false,true,true,false,true,false,false,true,false,false,false,true,false,false,false,false,true,false,false,false,true,false,false,true,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":15,"left_children":[1,3,5,7,11,9,13,15,27,17,29,25,21,19,23,-1,31,-1,-1,43,-1,33,45,-1,41,-1,-1,37,35,39,-1,47,57,49,-1,-1,59,-1,-1,-1,55,-1,-1,51,53,-1,-1,-1,61,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[1.4830469E0,1.5374919E0,1.286771E0,1.6355597E0,1.080421E0,1.0528343E0,8.800626E-1,1.6227047E0,9.755007E-1,1.0058506E0,8.0912244E-1,4.3297005E-1,7.74561E-1,5.949147E-1,7.146998E-1,0E0,9.9228716E-1,0E0,0E0,4.588499E-1,0E0,6.478387E-1,6.671761E-1,0E0,2.2489777E-1,0E0,0E0,4.749799E-1,7.560503E-1,6.725123E-1,0E0,4.4447994E-1,4.3417406E-1,3.4764016E-1,0E0,0E0,5.469712E-1,0E0,0E0,0E0,1.7368436E-1,0E0,0E0,3.5190728E-1,7.1306705E-2,0E0,0E0,0E0,1.5443754E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,13,13,12,12,14,14,11,11,8,8,10,10,16,16,21,21,28,28,27,27,29,29,24,24,19,19,22,22,31,31,33,33,43,43,44,44,40,40,32,32,36,36,48,48],"right_children":[2,4,6,8,12,10,14,16,28,18,30,26,22,20,24,-1,32,-1,-1,44,-1,34,46,-1,42,-1,-1,38,36,40,-1,48,58,50,-1,-1,60,-1,-1,-1,56,-1,-1,52,54,-1,-1,-1,62,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[1.7952664E0,2.6346316E0,-6.5128255E-1,2.136144E0,-4.125541E0,-3.2279758E0,4.8504663E-1,-9.99127E0,-5.0412595E-1,-1.2787838E0,7.3998946E-1,-2.1443331E0,1.2458334E0,6.647702E0,-1.77696E-1,9.987462E-2,6.5380344E0,-6.811813E-2,1.8228261E-1,-4.3961868E-1,-1.8645624E-2,5.954255E-1,-4.375813E0,-1.11821406E-1,-3.1438265E-2,-2.2008458E-4,1.9522381E-1,-2.7670827E0,-5.230856E0,-1.1830629E0,6.623818E-2,-8.410487E0,-1.2983434E-1,-4.356463E-1,5.9260186E-2,-7.117387E-2,-2.8781197E0,3.2858398E-2,-1.6346568E-1,4.647232E-2,4.622554E-1,1.5265906E-1,-7.950072E-3,8.449635E-2,-9.226288E-1,-6.969227E-2,1.5293838E-1,-3.191635E-2,8.3107815E0,2.503448E-3,-1.927057E-1,-4.8288498E-2,1.270997E-1,6.176983E-2,2.5125375E-1,-2.1010314E-1,-3.304957E-2,-8.854166E-2,1.0878457E-1,-2.0140395E-2,2.0131119E-1,-2.4775034E-1,-6.729467E-2],"split_indices":[133,56,84,3,64,106,280,61,446,55,475,76,50,165,264,0,98,0,0,343,0,473,46,0,283,0,0,90,90,274,0,86,323,434,0,0,9,0,0,0,324,0,0,442,384,0,0,0,106,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[6.061107E1,3.5971928E1,2.4639141E1,2.3272259E1,1.2699671E1,1.1402406E1,1.3236734E1,1.4712164E1,8.560095E0,4.2840934E0,7.118312E0,4.452E0,8.24767E0,9.192556E0,4.0441785E0,1.5744889E0,1.3137674E1,1.739377E0,2.5447164E0,7.949478E0,1.2430785E0,4.871131E0,3.3765392E0,1.8417568E0,2.2024217E0,1.3554024E0,3.096598E0,3.236884E0,5.3232107E0,5.37743E0,1.7408822E0,1.1110509E1,2.027166E0,3.5818934E0,1.2892373E0,1.2770158E0,4.046195E0,1.3353591E0,1.901525E0,1.0396403E0,4.3377895E0,1.0998238E0,1.1025978E0,2.3699908E0,5.5794873E0,1.221296E0,2.1552432E0,1.5201286E0,9.590381E0,1.0353597E0,2.5465338E0,1.2051322E0,1.1648586E0,1.0076045E0,4.5718827E0,3.29302E0,1.0447696E0,1.0217077E0,1.0054582E0,1.1214931E0,2.924702E0,8.176294E0,1.414086E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"63","size_leaf_vector":"0"}},{"base_weights":[2.7447762E-3,-1.3175389E-1,1.5482642E-1,-3.3446032E-1,8.831178E-2,-9.999443E-2,3.076592E-1,-5.174128E-1,-6.474583E-2,2.9371056E-1,-2.781281E-1,-2.6960397E-1,3.2677642E-1,4.5405632E-1,-1.632893E-1,4.178201E-2,-6.126831E-1,1.5455613E-1,-3.559821E-2,-1.7972054E-2,5.789943E-1,-2.2884925E-1,5.3369135E-1,-1.4108361E-1,6.556497E-2,-4.98494E-1,6.63419E-2,1.1279573E-1,-2.7211437E-1,6.7108124E-2,-4.304371E-1,-6.9268453E-1,-1.26961395E-2,6.670199E-1,-3.204466E-2,6.633387E-2,-4.471462E-1,5.617665E-4,-2.0033047E-1,1.3660937E-2,-6.0584635E-1,-2.0240212E-2,6.4534134E-1,-1.0543042E-1,1.0276284E-1,5.2887086E-2,-1.4999308E-1,-2.295943E-1,-5.5632096E-2,9.7730905E-3,-1.7752022E-1,2.221073E-1,2.9114189E-2,5.0025463E-2,7.4123955E-1,-3.8439136E-2,-2.1175711E-1,2.4414332E-1,3.5309594E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,true,false,false,true,false,true,true,true,false,false,false,true,false,true,false,false,true,true,true,false,false,false,true,false,false,true,false,false,true,false,false,false,false,true,false,false,false,false,false,true,false,false,false,false,false,false,false,false,false,false,false,true,false,false,false,false],"id":16,"left_children":[1,3,5,7,11,9,13,15,27,17,29,25,21,19,23,-1,31,-1,-1,43,33,45,41,-1,-1,37,-1,-1,35,-1,39,47,-1,53,-1,-1,49,-1,-1,-1,55,-1,51,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,57,-1,-1,-1,-1],"loss_changes":[1.1792334E0,1.394785E0,1.0841681E0,7.629119E-1,1.4080645E0,8.458082E-1,1.2567711E0,6.7802167E-1,8.0294997E-1,4.1553268E-1,7.0888907E-1,8.663919E-1,1.2411678E0,7.977073E-1,6.6581047E-1,0E0,3.5664487E-1,0E0,0E0,5.7286763E-1,7.1031547E-1,4.5513862E-1,5.3971434E-1,0E0,0E0,4.1610074E-1,0E0,0E0,5.990161E-1,0E0,5.69502E-1,1.2897038E-1,0E0,2.2986841E-1,0E0,0E0,3.38818E-1,0E0,0E0,0E0,1.3949692E-1,0E0,2.5788093E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,2.662139E-1,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,13,13,12,12,14,14,11,11,8,8,10,10,16,16,20,20,28,28,25,25,30,30,22,22,19,19,21,21,31,31,36,36,42,42,33,33,40,40,54,54],"right_children":[2,4,6,8,12,10,14,16,28,18,30,26,22,20,24,-1,32,-1,-1,44,34,46,42,-1,-1,38,-1,-1,36,-1,40,48,-1,54,-1,-1,50,-1,-1,-1,56,-1,52,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,58,-1,-1,-1,-1],"split_conditions":[6.417246E-1,8.1120896E-1,-2.7148666E0,1.468729E0,-1.8005326E0,-4.5544E0,4.5872655E0,-8.851955E0,-3.8082528E0,-2.6305366E0,-9.4535697E-1,2.461223E0,-6.3173084E0,-7.6605684E-1,-2.266486E-1,4.178201E-2,6.202408E0,1.5455613E-1,-3.559821E-2,-6.284797E-1,9.9786353E-1,-2.4284768E0,-7.5044215E-1,-1.4108361E-1,6.556497E-2,-2.890792E0,6.63419E-2,1.1279573E-1,-5.5992026E0,6.7108124E-2,-4.1217785E0,7.663866E-1,-1.26961395E-2,-6.910336E-1,-3.204466E-2,6.633387E-2,-5.0594707E0,5.617665E-4,-2.0033047E-1,1.3660937E-2,-8.557349E-1,-2.0240212E-2,3.6409216E0,-1.0543042E-1,1.0276284E-1,5.2887086E-2,-1.4999308E-1,-2.295943E-1,-5.5632096E-2,9.7730905E-3,-1.7752022E-1,2.221073E-1,2.9114189E-2,5.0025463E-2,1.0667093E0,-3.8439136E-2,-2.1175711E-1,2.4414332E-1,3.5309594E-2],"split_indices":[78,36,3,137,150,141,117,8,131,75,348,100,91,292,318,0,2,0,0,19,196,88,214,0,0,164,0,0,111,0,122,356,0,304,0,0,25,0,0,0,451,0,58,0,0,0,0,0,0,0,0,0,0,0,408,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[5.565348E1,2.9604965E1,2.6048515E1,1.5139698E1,1.4465265E1,9.895423E0,1.6153091E1,8.473554E0,6.666145E0,2.8804123E0,7.015011E0,5.73177E0,8.733496E0,1.23499155E1,3.8031762E0,1.0150526E0,7.458501E0,1.7117412E0,1.1686711E0,2.7638733E0,9.586042E0,2.3410637E0,6.3924317E0,1.979828E0,1.8233482E0,3.8979266E0,1.8338436E0,1.8727533E0,4.7933917E0,1.5242584E0,5.490752E0,6.352947E0,1.105553E0,8.519889E0,1.066154E0,1.1865385E0,3.6068532E0,1.2353194E0,2.6626072E0,1.677971E0,3.812781E0,1.0632548E0,5.329177E0,1.4488907E0,1.3149827E0,1.078811E0,1.2622527E0,5.112189E0,1.2407587E0,1.0147474E0,2.5921059E0,4.2435074E0,1.0856698E0,1.5201613E0,6.999728E0,1.0549278E0,2.7578533E0,5.996656E0,1.003072E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"59","size_leaf_vector":"0"}},{"base_weights":[6.0568965E-4,1.4899446E-1,-1.3294181E-1,2.9894137E-1,-1.6160677E-1,-2.3709677E-1,2.1988532E-1,-1.7832743E-1,4.4206452E-1,1.8243954E-1,-3.4019962E-1,-3.516432E-1,2.528974E-1,-6.0484592E-2,4.2317173E-1,6.761395E-2,-3.5543624E-1,-6.1820574E-2,3.9555186E-1,-6.3827224E-2,1.761268E-1,1.720322E-1,-1.0535657E-2,5.331334E-2,-4.9380165E-1,-4.6469606E-2,5.1297605E-1,9.922161E-2,-4.0166327E-1,-1.585497E-1,1.5109182E-2,-3.1455327E-2,-2.0473973E-1,-4.9098825E-1,5.9498504E-2,6.795484E-3,1.6110003E-1,6.072391E-1,3.7178264E-3,-1.6252708E-1,-6.4649653E-1,-9.789784E-2,1.1623414E-1,6.792553E-1,9.873093E-3,9.15892E-2,-1.00725465E-1,-2.2363023E-1,-1.5389932E-2,4.1049875E-2,2.2827618E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,true,true,true,true,false,false,true,true,false,true,false,true,false,true,false,false,false,false,false,false,false,true,false,true,false,false,false,false,false,false,false,false,false,false,true,false,false,true,false,false,false,false,false,false,false,false,false,false],"id":17,"left_children":[1,3,5,7,11,9,13,15,25,17,27,23,19,-1,21,-1,29,-1,35,-1,-1,-1,-1,-1,31,-1,37,-1,33,-1,-1,-1,-1,39,41,-1,-1,43,-1,45,47,-1,-1,49,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[1.0582751E0,1.2263352E0,1.0851116E0,1.2680392E0,7.887566E-1,1.0216334E0,6.787286E-1,4.1822332E-1,6.5857863E-1,4.9774176E-1,8.815923E-1,5.776374E-1,6.500325E-1,0E0,3.6411643E-1,0E0,2.8470388E-1,0E0,1.7713815E-1,0E0,0E0,0E0,0E0,0E0,3.244356E-1,0E0,5.804162E-1,0E0,7.329769E-1,0E0,0E0,0E0,0E0,6.3149333E-1,5.9005886E-1,0E0,0E0,4.1047287E-1,0E0,5.862503E-1,5.183017E-1,0E0,0E0,3.0552912E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,12,12,14,14,11,11,8,8,10,10,16,16,24,24,28,28,18,18,26,26,33,33,34,34,37,37,39,39,40,40,43,43],"right_children":[2,4,6,8,12,10,14,16,26,18,28,24,20,-1,22,-1,30,-1,36,-1,-1,-1,-1,-1,32,-1,38,-1,34,-1,-1,-1,-1,40,42,-1,-1,44,-1,46,48,-1,-1,50,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[-1.8251777E-1,4.9431257E0,6.1883984E0,-5.7621984E0,5.3205955E-1,-7.4188886E0,-3.4838343E0,-6.923653E-1,-1.2775596E0,-5.4560214E-1,-1.3765718E1,-7.790003E-1,-1.2511156E0,-6.0484592E-2,1.6098351E0,6.761395E-2,4.487216E-1,-6.1820574E-2,-1.1053629E1,-6.3827224E-2,1.761268E-1,1.720322E-1,-1.0535657E-2,5.331334E-2,-1.8841381E0,-4.6469606E-2,6.8340015E0,9.922161E-2,6.678671E0,-1.585497E-1,1.5109182E-2,-3.1455327E-2,-2.0473973E-1,-3.1105676E-1,-1.701993E-2,6.795484E-3,1.6110003E-1,1.0880716E0,3.7178264E-3,-8.7681794E-1,7.9052687E0,-9.789784E-2,1.1623414E-1,-6.4845586E0,9.873093E-3,9.15892E-2,-1.00725465E-1,-2.2363023E-1,-1.5389932E-2,4.1049875E-2,2.2827618E-1],"split_indices":[159,25,103,48,391,126,154,496,252,436,81,305,94,0,22,0,398,0,126,0,0,0,0,0,146,0,86,0,154,0,0,0,0,280,422,0,0,408,0,445,19,0,0,133,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[5.1402992E1,2.4293766E1,2.7109226E1,1.6409431E1,7.8843346E0,2.1183764E1,5.9254627E0,3.728876E0,1.2680555E1,4.027195E0,1.7156569E1,5.509887E0,2.3744473E0,1.9306374E0,3.9948254E0,1.0539409E0,2.674935E0,1.4382633E0,2.5889316E0,1.1439136E0,1.2305337E0,2.819183E0,1.1756423E0,1.1134925E0,4.3963947E0,1.2206035E0,1.1459951E1,1.0721308E0,1.6084438E1,1.6632442E0,1.0116909E0,1.9441284E0,2.4522665E0,1.3420308E1,2.66413E0,1.0310123E0,1.5579193E0,9.464636E0,1.9953153E0,4.969331E0,8.450977E0,1.2263281E0,1.4378018E0,8.247742E0,1.2168943E0,1.136788E0,3.8325427E0,7.0299892E0,1.4209878E0,1.4293703E0,6.8183713E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"51","size_leaf_vector":"0"}},{"base_weights":[4.0156594E-3,-1.0737499E-1,1.6172838E-1,-2.3961583E-1,1.3167249E-1,3.845929E-3,4.1155234E-1,1.8567161E-1,-3.3891493E-1,-2.0112552E-1,2.9141933E-1,-2.892963E-1,2.719086E-1,6.1819416E-1,-6.41972E-2,-6.5963395E-2,1.3472787E-1,8.152635E-2,-3.448318E-1,2.5471054E-2,2.214736E-1,-4.728458E-2,4.1479918E-1,6.986373E-2,-1.0083104E-1,-1.439596E-1,1.5976217E-2,-4.750084E-1,1.338545E-1,4.7273025E-1,-2.876643E-2,-5.219766E-1,2.42733E-2,-5.407731E-1,4.186583E-2,1.7456323E-2,1.7974658E-1,5.4486245E-1,-1.8230526E-2,-6.2167645E-2,1.1781074E-1,-2.495859E-2,-1.8357377E-1,-6.630094E-1,-1.0965364E-1,3.2341205E-2,1.940659E-1,-4.2709373E-2,-2.2643282E-1,7.293911E-2,-1.2222007E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,true,true,true,true,false,true,true,false,false,false,true,false,false,false,true,false,false,false,true,false,false,false,false,true,false,false,false,true,false,true,false,false,false,false,false,false,false,false,false,false,true,false,false,false,false,false,false],"id":18,"left_children":[1,3,5,7,11,9,13,15,27,17,29,25,21,19,23,-1,-1,-1,31,-1,-1,-1,37,-1,-1,-1,-1,33,39,35,-1,41,-1,43,-1,-1,-1,45,-1,-1,-1,-1,-1,47,49,-1,-1,-1,-1,-1,-1],"loss_changes":[8.75008E-1,9.550066E-1,8.112104E-1,8.7217176E-1,7.309906E-1,8.623997E-1,8.232627E-1,5.552675E-1,1.0859034E0,6.7495584E-1,4.7250265E-1,2.3743168E-1,6.057331E-1,2.9910564E-1,3.4641412E-1,0E0,0E0,0E0,5.702014E-1,0E0,0E0,0E0,4.5877337E-1,0E0,0E0,0E0,0E0,5.8299327E-1,4.638155E-1,1.9112515E-1,0E0,1.4947295E-1,0E0,5.3589654E-1,0E0,0E0,0E0,1.7569673E-1,0E0,0E0,0E0,0E0,0E0,3.129635E-1,4.9287522E-1,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,13,13,12,12,14,14,11,11,8,8,10,10,18,18,27,27,29,29,22,22,28,28,31,31,33,33,37,37,43,43,44,44],"right_children":[2,4,6,8,12,10,14,16,28,18,30,26,22,20,24,-1,-1,-1,32,-1,-1,-1,38,-1,-1,-1,-1,34,40,36,-1,42,-1,44,-1,-1,-1,46,-1,-1,-1,-1,-1,48,50,-1,-1,-1,-1,-1,-1],"split_conditions":[1.7952664E0,2.6346316E0,3.5089332E-1,-7.366765E0,-5.6630373E0,4.6738672E-1,4.935152E-1,-2.456881E-1,7.974899E-1,-5.8833914E0,4.458763E0,-3.179036E-2,-6.4332205E-1,-5.890598E-1,2.157022E-1,-6.5963395E-2,1.3472787E-1,8.152635E-2,4.225803E-1,2.5471054E-2,2.214736E-1,-4.728458E-2,5.2369815E-1,6.986373E-2,-1.0083104E-1,-1.439596E-1,1.5976217E-2,1.0145633E1,3.586452E-1,-1.0248486E-1,-2.876643E-2,-6.6006975E0,2.42733E-2,3.6474152E0,4.186583E-2,1.7456323E-2,1.7974658E-1,-4.382659E0,-1.8230526E-2,-6.2167645E-2,1.1781074E-1,-2.495859E-2,-1.8357377E-1,-6.3737373E0,2.2674498E-1,3.2341205E-2,1.940659E-1,-4.2709373E-2,-2.2643282E-1,7.293911E-2,-1.2222007E-1],"split_indices":[133,56,264,61,151,177,168,234,191,26,128,445,389,436,290,0,0,0,294,0,0,0,315,0,0,0,0,17,3,218,0,12,0,3,0,0,0,94,0,0,0,0,0,127,180,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[4.7819214E1,2.8212341E1,1.9606873E1,1.816249E1,1.004985E1,1.2636388E1,6.9704843E0,3.2733335E0,1.4889157E1,7.553295E0,5.0830927E0,2.2456894E0,7.804161E0,4.650938E0,2.3195462E0,1.3540335E0,1.9193001E0,1.5527552E0,6.00054E0,1.1683087E0,3.4826293E0,1.9224112E0,5.881749E0,1.1770594E0,1.1424868E0,1.1853758E0,1.0603136E0,1.1557446E1,3.3317115E0,3.3120575E0,1.7710352E0,4.077796E0,1.922744E0,1.0548688E1,1.0087574E0,1.1152973E0,2.1967602E0,4.5041866E0,1.3775628E0,1.523565E0,1.8081465E0,1.0210972E0,3.0566988E0,7.799421E0,2.749267E0,1.2416699E0,3.2625167E0,1.5510124E0,6.2484083E0,1.3422906E0,1.4069765E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"51","size_leaf_vector":"0"}},{"base_weights":[6.2180334E-3,1.2990148E-1,-1.4327104E-1,-1.1178314E-1,2.993737E-1,3.0355072E-1,-2.2760841E-1,-3.6824948E-1,2.1957707E-1,-8.123679E-3,1.4664929E-1,-8.853534E-2,3.917994E-1,-4.3576178E-1,6.539162E-2,-5.489781E-1,6.565106E-2,5.120137E-1,-5.302979E-2,4.3218574E-1,-1.9157092E-1,4.282564E-1,-9.053406E-2,1.9073097E-2,-5.410598E-1,-2.8821474E-2,-1.9518653E-1,-9.1961585E-2,7.256186E-2,1.6071077E-1,1.0944231E-2,2.8651727E-2,-6.321286E-1,-4.1865242E-1,4.8626363E-2,1.7807003E-2,6.4818853E-1,1.8799543E-2,1.8250206E-1,-4.235004E-2,-2.1317548E-1,-8.073328E-2,9.498838E-2,-8.595691E-3,-1.696E-1,2.2657072E-2,7.147904E-1,7.3195085E-2,2.3833443E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,true,false,false,false,true,false,false,false,false,true,true,true,true,false,false,true,true,false,true,false,false,true,false,false,false,false,false,false,false,true,true,false,true,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":19,"left_children":[1,3,5,7,11,9,13,15,21,-1,-1,-1,17,23,19,25,-1,35,27,37,33,29,-1,-1,31,-1,-1,-1,-1,-1,-1,-1,39,43,-1,41,45,-1,-1,-1,-1,-1,-1,-1,-1,-1,47,-1,-1],"loss_changes":[8.627211E-1,1.0782347E0,8.659557E-1,1.0362376E0,9.5442486E-1,2.4516943E-1,1.148253E0,8.32166E-1,7.310264E-1,0E0,0E0,0E0,7.566216E-1,6.056578E-1,8.8349855E-1,1.8682683E-1,0E0,7.175014E-1,3.588595E-1,2.0967674E-1,5.246959E-1,1.6414076E-1,0E0,0E0,5.8662033E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,2.1064758E-1,1.9793749E-1,0E0,3.7400368E-1,2.688613E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,8.611488E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,12,12,14,14,8,8,13,13,15,15,18,18,21,21,24,24,20,20,17,17,19,19,32,32,35,35,33,33,36,36,46,46],"right_children":[2,4,6,8,12,10,14,16,22,-1,-1,-1,18,24,20,26,-1,36,28,38,34,30,-1,-1,32,-1,-1,-1,-1,-1,-1,-1,40,44,-1,42,46,-1,-1,-1,-1,-1,-1,-1,-1,-1,48,-1,-1],"split_conditions":[1.139224E0,-9.9897707E-1,-1.1460266E0,2.7036667E-3,-8.4463825E0,-2.0875206E0,4.15802E-4,6.020973E0,5.129281E-1,-8.123679E-3,1.4664929E-1,-8.853534E-2,6.6764507E0,-7.3205347E0,-2.4396048E0,-6.4899576E-1,6.565106E-2,-6.3335733E0,1.1829271E0,-3.577193E-1,7.341981E-1,2.6513333E0,-9.053406E-2,1.9073097E-2,-1.0091861E1,-2.8821474E-2,-1.9518653E-1,-9.1961585E-2,7.256186E-2,1.6071077E-1,1.0944231E-2,2.8651727E-2,-3.3271177E0,-2.7174459E0,4.8626363E-2,2.7143128E0,-6.592561E0,1.8799543E-2,1.8250206E-1,-4.235004E-2,-2.1317548E-1,-8.073328E-2,9.498838E-2,-8.595691E-3,-1.696E-1,2.2657072E-2,-3.6865697E0,7.3195085E-2,2.3833443E-1],"split_indices":[132,161,260,156,138,126,144,153,403,0,0,0,68,162,122,255,0,151,154,376,112,27,0,0,161,0,0,0,0,0,0,0,69,63,0,133,79,0,0,0,0,0,0,0,0,0,33,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[4.4653904E1,2.4507803E1,2.0146101E1,1.0242029E1,1.4265774E1,2.7860878E0,1.7360014E1,5.710718E0,4.531311E0,1.3072208E0,1.4788669E0,1.6244059E0,1.2641368E1,9.86469E0,7.495324E0,4.4161587E0,1.2945592E0,9.832987E0,2.8083808E0,2.806855E0,4.688469E0,3.363575E0,1.1677359E0,1.7869459E0,8.077744E0,1.165971E0,3.2501876E0,1.4889604E0,1.3194205E0,2.3513396E0,1.0122353E0,1.0048916E0,7.072852E0,2.7432654E0,1.9452035E0,2.3683865E0,7.4646006E0,1.343766E0,1.463089E0,1.3602185E0,5.7126336E0,1.2589799E0,1.1094066E0,1.076464E0,1.6668013E0,1.0000234E0,6.464577E0,1.5234314E0,4.941146E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"49","size_leaf_vector":"0"}},{"base_weights":[8.719827E-3,1.21230505E-1,-1.5194893E-1,1.9987947E-1,-3.3442757E-1,3.45028E-1,-2.6499328E-1,4.184539E-1,2.6955778E-2,-1.8366598E-2,1.5482314E-1,-1.4562252E-1,-9.889264E-3,-3.4879348E-1,6.367243E-2,5.179771E-1,-2.13865E-2,3.2306504E-1,-1.4869042E-1,-4.291007E-1,3.707133E-2,5.933931E-1,-1.961992E-3,2.5201514E-1,-3.2390836E-1,-2.6100034E-2,4.5334628E-1,1.7460542E-2,-5.060489E-1,2.0225735E-1,4.232775E-2,5.913945E-2,-4.846094E-1,-1.3199797E-2,-5.669352E-1,-2.5235545E-2,1.4450563E-1,1.7260635E-2,1.884797E-1,6.484532E-3,-1.8624665E-1,-1.1836312E-2,-1.8940502E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,true,false,false,false,true,true,false,false,false,false,true,false,true,false,true,true,false,false,true,false,true,true,false,true,false,false,false,false,false,true,false,false,false,false,false,false,false,false,false,false],"id":20,"left_children":[1,3,5,7,11,9,13,15,17,-1,-1,-1,-1,19,-1,21,-1,25,23,27,-1,29,-1,35,31,-1,37,-1,33,-1,-1,-1,39,-1,41,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[7.865115E-1,9.8235524E-1,1.0995077E0,8.3581966E-1,1.7218068E-1,2.928422E-1,6.8127275E-1,5.1606965E-1,7.544183E-1,0E0,0E0,0E0,0E0,5.5733466E-1,0E0,3.318529E-1,0E0,3.2642978E-1,7.282436E-1,4.638996E-1,0E0,1.7018557E-1,0E0,3.002446E-1,6.6869813E-1,0E0,2.572512E-1,0E0,2.59454E-1,0E0,0E0,0E0,3.9410543E-1,0E0,2.7572727E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,8,8,13,13,15,15,18,18,17,17,19,19,21,21,24,24,28,28,23,23,26,26,32,32,34,34],"right_children":[2,4,6,8,12,10,14,16,18,-1,-1,-1,-1,20,-1,22,-1,26,24,28,-1,30,-1,36,32,-1,38,-1,34,-1,-1,-1,40,-1,42,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[2.246585E-1,8.174833E0,-7.502468E0,-1.601979E0,2.8415278E-1,-4.2137197E-1,1.0938679E0,8.115318E-1,-4.0582716E-1,-1.8366598E-2,1.5482314E-1,-1.4562252E-1,-9.889264E-3,6.9469776E0,6.367243E-2,6.7967625E0,-2.13865E-2,-7.8923607E-1,-4.227988E0,-1.083694E0,3.707133E-2,6.204967E0,-1.961992E-3,-2.3923676E-1,-6.5369806E0,-2.6100034E-2,-1.9298775E-1,1.7460542E-2,-1.0467466E0,2.0225735E-1,4.232775E-2,5.913945E-2,-5.2442226E0,-1.3199797E-2,-1.16739E0,-2.5235545E-2,1.4450563E-1,1.7260635E-2,1.884797E-1,6.484532E-3,-1.8624665E-1,-1.1836312E-2,-1.8940502E-1],"split_indices":[226,147,159,61,486,466,264,242,280,0,0,0,0,158,0,91,0,393,27,414,0,132,0,225,22,0,434,0,322,0,0,0,64,0,380,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[4.1490047E1,2.4546478E1,1.694357E1,2.1411995E1,3.134483E0,2.7595608E0,1.4184008E1,8.830466E0,1.2581529E1,1.0078785E0,1.7516823E0,1.6819462E0,1.452537E0,1.2294319E1,1.889689E0,7.2911406E0,1.5393255E0,4.3719234E0,8.209605E0,1.0586147E1,1.7081718E0,6.2597847E0,1.0313556E0,2.3643296E0,5.845276E0,1.1341999E0,3.2377236E0,1.4768817E0,9.109265E0,4.894805E0,1.3649799E0,1.3244206E0,4.5208554E0,1.2611765E0,7.8480887E0,1.216973E0,1.1473566E0,1.3996025E0,1.838121E0,1.1369138E0,3.3839417E0,1.0295783E0,6.8185105E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"43","size_leaf_vector":"0"}},{"base_weights":[1.3462192E-2,1.4246468E-1,-1.256205E-1,2.8545102E-1,-1.4099887E-1,-2.4300435E-1,1.8430611E-1,-1.01022564E-1,3.8158166E-1,-3.9606762E-1,4.290874E-2,-4.230146E-1,1.2772907E-1,-8.219956E-2,3.7011388E-1,-1.00580774E-1,7.464965E-2,-6.267664E-1,-9.644706E-2,-5.3927008E-2,1.1068572E-1,1.5950988E-1,-8.707811E-3,-1.568121E-1,-1.958359E-2,-3.849054E-2,4.7470406E-1,3.6165324E-1,-2.4116915E-1,-2.1058601E-1,-4.8783008E-2,-1.906106E-2,5.516716E-1,-1.319192E-1,2.5708152E-2,1.8888089E-1,-1.1559808E-1,1.6254297E-1,-2.4890418E-3,-9.045267E-4,6.3851625E-1,-4.7859363E-2,1.23469085E-1,2.1062575E-1,3.684302E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,true,true,true,true,true,true,true,true,true,false,true,false,false,true,true,false,false,false,false,false,false,false,false,false,true,false,false,false,true,false,false,true,false,false,false,false,true,false,false,false,false],"id":21,"left_children":[1,3,5,7,11,9,13,15,25,17,27,23,19,-1,21,-1,-1,29,35,-1,-1,-1,-1,-1,-1,-1,31,37,33,-1,-1,-1,39,-1,-1,41,-1,-1,-1,-1,43,-1,-1,-1,-1],"loss_changes":[7.356707E-1,9.004588E-1,7.644607E-1,5.749438E-1,6.4405364E-1,6.6376764E-1,6.1259294E-1,3.8391E-1,6.1982536E-1,6.116667E-1,6.3553E-1,1.0962796E-1,4.1899157E-1,0E0,3.1834137E-1,0E0,0E0,7.122898E-2,5.188527E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,4.562087E-1,2.213493E-1,2.8442648E-1,0E0,0E0,0E0,4.353552E-1,0E0,0E0,3.2895905E-1,0E0,0E0,0E0,0E0,1.6897154E-1,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,12,12,14,14,11,11,8,8,10,10,17,17,26,26,28,28,18,18,27,27,32,32,35,35,40,40],"right_children":[2,4,6,8,12,10,14,16,26,18,28,24,20,-1,22,-1,-1,30,36,-1,-1,-1,-1,-1,-1,-1,32,38,34,-1,-1,-1,40,-1,-1,42,-1,-1,-1,-1,44,-1,-1,-1,-1],"split_conditions":[-1.420579E-1,3.810327E-1,4.499092E0,-6.086714E0,-6.73686E-1,3.5339298E0,-4.7594423E0,2.6132202E0,-8.4401786E-1,-4.9670494E-1,-3.12778E-1,3.7672886E-1,-2.7847767E-2,-8.219956E-2,4.8434323E-1,-1.00580774E-1,7.464965E-2,3.6466278E1,2.3010038E-1,-5.3927008E-2,1.1068572E-1,1.5950988E-1,-8.707811E-3,-1.568121E-1,-1.958359E-2,-3.849054E-2,-1.2458792E0,-1.6264145E0,-7.312435E-2,-2.1058601E-1,-4.8783008E-2,-1.906106E-2,-6.9924836E0,-1.319192E-1,2.5708152E-2,1.4281636E-1,-1.1559808E-1,1.6254297E-1,-2.4890418E-3,-9.045267E-4,9.9743766E-1,-4.7859363E-2,1.23469085E-1,2.1062575E-1,3.684302E-2],"split_indices":[147,238,5,66,11,36,160,78,409,151,26,469,19,0,378,0,0,166,418,0,0,0,0,0,0,0,318,162,47,0,0,0,34,0,0,298,0,0,0,0,350,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[3.900522E1,2.0223307E1,1.8781912E1,1.3437881E1,6.7854247E0,1.377906E1,5.002852E0,2.666575E0,1.0771306E1,8.723612E0,5.0554485E0,3.030714E0,3.754711E0,1.3062594E0,3.6965926E0,1.6221486E0,1.0444263E0,4.311794E0,4.411818E0,1.7626654E0,1.9920455E0,2.3949287E0,1.3016639E0,2.0095484E0,1.0211655E0,1.6050757E0,9.16623E0,2.253675E0,2.8017735E0,3.269809E0,1.0419847E0,1.1686007E0,7.997629E0,1.5281518E0,1.2736218E0,2.3937054E0,2.0181127E0,1.2031982E0,1.0504768E0,1.2133172E0,6.7843122E0,1.0439392E0,1.3497661E0,5.7180576E0,1.0662544E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"45","size_leaf_vector":"0"}},{"base_weights":[7.4211517E-3,1.0777233E-1,-1.6918279E-1,-7.135587E-2,2.7197978E-1,2.4819176E-1,-2.824065E-1,-1.809703E-1,3.1426603E-1,1.3839328E-1,-2.153463E-2,-2.272567E-2,4.288664E-1,-3.840478E-1,4.2192858E-2,-2.9324934E-1,9.014594E-2,-3.532064E-1,3.2343173E-1,-4.623574E-1,4.016905E-2,1.2152868E-2,1.2728457E-1,5.551365E-1,7.506408E-2,-4.2709666E-1,1.1215255E-1,1.3109502E-1,1.2747818E-2,-7.860094E-2,1.0031352E-1,3.375758E-2,1.9434191E-1,-1.5505406E-2,-1.4709331E-1,-5.516319E-1,2.3090664E-2,-5.316137E-1,2.7427524E-3,-4.0282886E-2,9.064278E-2,-6.3660085E-1,-3.1021686E-2,-2.8244305E-2,-1.9114827E-1,-3.349178E-2,-2.1645717E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,true,true,true,false,true,true,false,false,false,true,true,false,true,false,true,false,true,false,false,false,true,true,true,false,false,false,false,false,false,false,false,false,true,false,false,false,false,false,false,false,false,false,false,false],"id":22,"left_children":[1,3,5,7,11,9,13,15,21,-1,-1,17,23,19,-1,25,-1,33,27,35,-1,-1,-1,31,29,37,39,-1,-1,-1,-1,-1,-1,-1,-1,41,-1,43,-1,-1,-1,45,-1,-1,-1,-1,-1],"loss_changes":[6.841059E-1,7.4209476E-1,7.3182625E-1,5.8706087E-1,6.1202514E-1,2.7317375E-1,5.4776025E-1,6.503868E-1,7.743144E-2,0E0,0E0,7.405853E-1,3.4035277E-1,4.4829655E-1,0E0,5.3302693E-1,0E0,1.1021212E-1,8.3391994E-2,4.5170534E-1,0E0,0E0,0E0,1.791817E-1,3.852733E-1,3.2079887E-1,1.815491E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,2.165761E-1,0E0,1.9890344E-1,0E0,0E0,0E0,1.886971E-1,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,11,11,13,13,8,8,12,12,15,15,18,18,24,24,23,23,17,17,19,19,25,25,26,26,35,35,37,37,41,41],"right_children":[2,4,6,8,12,10,14,16,22,-1,-1,18,24,20,-1,26,-1,34,28,36,-1,-1,-1,32,30,38,40,-1,-1,-1,-1,-1,-1,-1,-1,42,-1,44,-1,-1,-1,46,-1,-1,-1,-1,-1],"split_conditions":[2.2912774E0,-1.9463277E-1,-8.009977E0,7.6190567E0,-3.230455E0,-4.7899175E-1,9.2635804E-1,6.779244E0,-1.0177867E0,1.3839328E-1,-2.153463E-2,-1.2920866E0,3.895272E0,1.20451E0,4.2192858E-2,6.807881E-1,9.014594E-2,5.8873696E-3,4.2713615E-1,7.7933846E0,4.016905E-2,1.2152868E-2,1.2728457E-1,-9.270571E-1,-2.869094E0,6.1171374E0,-2.3748393E0,1.3109502E-1,1.2747818E-2,-7.860094E-2,1.0031352E-1,3.375758E-2,1.9434191E-1,-1.5505406E-2,-1.4709331E-1,5.2095566E0,2.3090664E-2,-4.9787357E-1,2.7427524E-3,-4.0282886E-2,9.064278E-2,-7.0061216E0,-3.1021686E-2,-2.8244305E-2,-1.9114827E-1,-3.349178E-2,-2.1645717E-1],"split_indices":[81,73,61,161,108,125,267,55,126,0,0,112,114,429,0,348,0,249,197,71,0,0,0,461,59,146,152,0,0,0,0,0,0,0,0,70,0,496,0,0,0,164,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[3.6572342E1,2.3569347E1,1.3002995E1,1.1542973E1,1.2026375E1,2.5203235E0,1.0482671E1,9.401319E0,2.1416538E0,1.2476257E0,1.2726978E0,4.475141E0,7.551234E0,8.526336E0,1.9563352E0,7.940461E0,1.4608581E0,2.2789938E0,2.1961472E0,7.499749E0,1.0265864E0,1.0061728E0,1.135481E0,5.145702E0,2.405532E0,5.9293237E0,2.0111372E0,1.1684408E0,1.0277064E0,1.0414475E0,1.3640844E0,1.2741834E0,3.8715184E0,1.1427718E0,1.136222E0,6.4150505E0,1.0846986E0,4.6069345E0,1.322389E0,1.0031872E0,1.0079501E0,5.0394673E0,1.375583E0,1.2632171E0,3.3437176E0,1.0240122E0,4.0154552E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"47","size_leaf_vector":"0"}},{"base_weights":[1.0930176E-2,-5.5583954E-2,2.6607403E-1,7.613612E-2,-2.2108904E-1,-8.337541E-2,4.1419828E-1,-2.9811552E-1,1.7013836E-1,-8.975988E-2,5.111616E-2,-4.368366E-1,1.5469143E-2,1.6912203E-1,1.0746047E-2,-1.4710104E-1,3.0921306E-2,-2.043139E-1,-1.3166802E-1,-1.0332405E-1,3.2294607E-1,-2.1778762E-1,2.5058433E-1,-1.1783511E-1,6.1710406E-2,-4.5245465E-2,4.273617E-1,1.3612904E-1,-1.8929206E-2,7.111689E-2,-3.246583E-1,-1.4219767E-1,3.820511E-2,8.8707305E-2,5.8626133E-1,-2.5019897E-3,-1.4896569E-1,-7.059341E-2,1.05418384E-1,1.9650362E-1,6.133413E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,true,true,false,true,true,true,false,false,true,true,false,false,false,false,false,true,false,true,true,true,false,false,false,true,false,false,false,true,false,false,false,true,false,false,false,false,false,false],"id":23,"left_children":[1,3,5,7,11,9,13,15,19,-1,-1,17,21,-1,-1,-1,-1,-1,23,29,25,31,27,-1,-1,-1,33,-1,-1,-1,35,-1,-1,37,39,-1,-1,-1,-1,-1,-1],"loss_changes":[6.1503863E-1,6.464931E-1,4.2206305E-1,6.4269966E-1,6.731018E-1,2.1863547E-1,2.8680325E-1,3.3906022E-1,6.2196314E-1,0E0,0E0,4.2176592E-1,4.4908223E-1,0E0,0E0,0E0,0E0,0E0,4.5873407E-1,5.1444477E-1,5.2345175E-1,4.269804E-1,2.8659117E-1,0E0,0E0,0E0,3.7116134E-1,0E0,0E0,0E0,2.1013471E-1,0E0,0E0,3.9673644E-1,1.9054413E-3,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,11,11,8,8,12,12,18,18,20,20,22,22,19,19,21,21,26,26,30,30,33,33,34,34],"right_children":[2,4,6,8,12,10,14,16,20,-1,-1,18,22,-1,-1,-1,-1,-1,24,30,26,32,28,-1,-1,-1,34,-1,-1,-1,36,-1,-1,38,40,-1,-1,-1,-1,-1,-1],"split_conditions":[6.313111E0,1.8050769E-1,-3.8044987E0,-9.166804E-1,-4.615353E-2,1.1496031E-1,2.3426156E0,1.7805247E0,-3.1832427E-1,-8.975988E-2,5.111616E-2,-1.913661E-1,1.3532506E-1,1.6912203E-1,1.0746047E-2,-1.4710104E-1,3.0921306E-2,-2.043139E-1,4.0081795E-2,-4.6043244E-1,-9.9250686E-1,-7.277336E-1,-1.0591057E-1,-1.1783511E-1,6.1710406E-2,-4.5245465E-2,-5.935863E-1,1.3612904E-1,-1.8929206E-2,7.111689E-2,-2.860284E-1,-1.4219767E-1,3.820511E-2,-1.3809862E0,4.4244272E-1,-2.5019897E-3,-1.4896569E-1,-7.059341E-2,1.05418384E-1,1.9650362E-1,6.133413E-2],"split_indices":[154,280,70,374,304,362,106,44,436,0,0,476,313,0,0,0,0,0,454,257,393,92,435,0,0,0,102,0,0,0,69,0,0,10,232,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[3.4369797E1,2.788304E1,6.4867573E0,1.5826902E1,1.2056137E1,2.0611923E0,4.4255652E0,2.7413573E0,1.30855465E1,1.0170996E0,1.0440927E0,5.8626256E0,6.1935115E0,2.8208356E0,1.6047297E0,1.7031212E0,1.0382361E0,2.5723836E0,3.290242E0,4.8069425E0,8.278604E0,3.1460385E0,3.047473E0,1.7621367E0,1.5281053E0,1.4148076E0,6.863796E0,1.5784806E0,1.4689924E0,1.8661062E0,2.9408364E0,1.5913777E0,1.5546608E0,2.6896806E0,4.1741157E0,1.4046061E0,1.5362303E0,1.2509177E0,1.4387629E0,2.930851E0,1.2432647E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"41","size_leaf_vector":"0"}},{"base_weights":[6.35554E-3,7.703093E-2,-2.1820591E-1,-1.4811961E-1,2.0146461E-1,4.7868367E-2,-3.4977132E-1,-3.1365684E-1,1.5128532E-1,-2.0097369E-1,2.793781E-1,-4.5886135E-1,7.65253E-3,-4.6764973E-1,2.169976E-2,3.0232336E-2,-1.2060245E-1,-2.470944E-2,-1.6087942E-1,1.0225773E-1,-4.4580676E-2,4.4488248E-1,3.9312743E-2,-1.80826E-2,-1.6783383E-1,5.3044564E-1,-1.0951911E-2,-1.9723365E-1,3.2942757E-1,6.1005974E-1,3.0310541E-2,-3.6484113E-1,3.4641713E-2,6.808517E-4,1.4305307E-1,4.377257E-2,2.0449089E-1,-1.3834083E-1,-3.053833E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,true,false,true,true,true,false,true,false,false,false,false,false,false,false,false,false,false,true,false,false,false,true,false,false,false,true,false,false,false,false,false,false,false,false,false],"id":24,"left_children":[1,3,5,7,9,-1,11,13,19,15,21,17,-1,23,-1,-1,-1,-1,-1,-1,-1,25,27,-1,-1,29,-1,31,33,35,-1,37,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[5.461968E-1,7.613456E-1,4.6563494E-1,5.36639E-1,5.967107E-1,0E0,2.7490777E-1,4.309625E-1,2.8878576E-1,2.4542373E-1,5.7300353E-1,1.0252607E-1,0E0,1.498872E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,3.777175E-1,5.564557E-1,0E0,0E0,1.9310975E-1,0E0,2.8168356E-1,1.721136E-1,1.0278773E-1,0E0,3.4522682E-2,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,6,6,7,7,9,9,11,11,8,8,10,10,13,13,21,21,22,22,25,25,27,27,28,28,29,29,31,31],"right_children":[2,4,6,8,10,-1,12,14,20,16,22,18,-1,24,-1,-1,-1,-1,-1,-1,-1,26,28,-1,-1,30,-1,32,34,36,-1,38,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[5.84927E0,-3.84905E0,-6.0129433E0,2.776455E0,-7.0387616E0,4.7868367E-2,4.6195297E0,3.705486E0,3.533002E-1,1.4749567E-1,2.3565753E-1,-5.8297706E-1,7.65253E-3,-6.6685116E-1,2.169976E-2,3.0232336E-2,-1.2060245E-1,-2.470944E-2,-1.6087942E-1,1.0225773E-1,-4.4580676E-2,1.1167015E0,3.093292E-2,-1.80826E-2,-1.6783383E-1,8.716961E-1,-1.0951911E-2,1.7007246E0,-2.1503055E-1,-5.689049E0,3.0310541E-2,-1.7442216E-1,3.4641713E-2,6.808517E-4,1.4305307E-1,4.377257E-2,2.0449089E-1,-1.3834083E-1,-3.053833E-2],"split_indices":[164,154,143,4,3,0,39,114,345,436,290,211,0,363,0,0,0,0,0,0,0,403,252,0,0,295,0,160,462,40,0,297,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[3.2355762E1,2.5109966E1,7.2457957E0,8.870078E0,1.6239887E1,1.816692E0,5.429104E0,5.6813374E0,3.1887407E0,2.3779364E0,1.3861951E1,4.0338006E0,1.3953034E0,3.9098976E0,1.7714397E0,1.1502131E0,1.2277232E0,1.0398719E0,2.9939287E0,1.8699989E0,1.3187418E0,7.7001724E0,6.1617785E0,1.0236697E0,2.8862278E0,6.4515514E0,1.2486207E0,3.5706148E0,2.5911634E0,5.0876026E0,1.363949E0,2.2160892E0,1.3545257E0,1.1202981E0,1.4708652E0,1.0856972E0,4.0019054E0,1.070974E0,1.1451153E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"39","size_leaf_vector":"0"}},{"base_weights":[3.4837297E-3,-7.4649654E-2,2.1338029E-1,-2.678137E-1,7.309933E-2,3.5937625E-1,-1.5667169E-1,-4.047186E-1,2.094956E-1,4.2807057E-2,1.5471053E-1,2.184259E-1,-1.6823877E-1,-1.06856935E-1,3.6392353E-2,4.1520802E-4,-5.124116E-1,-6.2441643E-2,7.992377E-2,6.73727E-2,-1.2219569E-1,-3.9484766E-3,9.825165E-2,-9.663943E-2,3.68426E-1,-2.2546317E-2,-5.9341055E-1,-1.0005823E-1,5.5755205E-2,-5.642681E-2,1.0140915E-1,-7.041183E-3,4.8427024E-1,-1.9598629E-1,-5.6590416E-2,1.7405817E-1,5.3803843E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,true,false,true,false,true,true,false,false,false,false,false,false,false,false,false,false,false,false,false,true,false,false,true,false,false,false,false,false,true,false,false,false,false],"id":25,"left_children":[1,3,5,7,11,9,13,15,21,17,-1,23,19,-1,-1,-1,25,-1,-1,29,-1,-1,-1,27,31,-1,33,-1,-1,-1,-1,-1,35,-1,-1,-1,-1],"loss_changes":[5.368897E-1,7.0138574E-1,5.3585553E-1,7.753664E-1,5.390506E-1,3.0794322E-1,2.2110872E-1,3.7479448E-1,8.351821E-2,2.4034266E-1,0E0,4.7862032E-1,3.6247408E-1,0E0,0E0,0E0,1.9494498E-1,0E0,0E0,3.300308E-1,0E0,0E0,0E0,3.1732494E-1,3.091098E-1,0E0,1.15561485E-2,0E0,0E0,0E0,0E0,0E0,4.535854E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,12,12,8,8,11,11,16,16,23,23,19,19,24,24,26,26,32,32],"right_children":[2,4,6,8,12,10,14,16,22,18,-1,24,20,-1,-1,-1,26,-1,-1,30,-1,-1,-1,28,32,-1,34,-1,-1,-1,-1,-1,36,-1,-1,-1,-1],"split_conditions":[5.2532053E0,-1.5896711E0,4.694896E0,6.4292207E0,4.1072986E-1,-2.9974685E0,1.07464194E-1,-7.071446E-1,-1.9650005E-1,1.2218094E-1,1.5471053E-1,-1.9350641E0,4.3438196E-1,-1.06856935E-1,3.6392353E-2,4.1520802E-4,-7.294865E0,-6.2441643E-2,7.992377E-2,-7.295811E-1,-1.2219569E-1,-3.9484766E-3,9.825165E-2,3.9781466E-2,-5.749799E0,-2.2546317E-2,5.914714E-1,-1.0005823E-1,5.5755205E-2,-5.642681E-2,1.0140915E-1,-7.041183E-3,2.97184E-1,-1.9598629E-1,-5.6590416E-2,1.7405817E-1,5.3803843E-2],"split_indices":[16,56,104,45,238,103,350,427,243,4,0,150,281,0,0,0,126,0,0,126,0,0,0,264,101,0,300,0,0,0,0,0,81,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[3.0766052E1,2.2889757E1,7.876294E0,9.568051E0,1.3321705E1,5.6686864E0,2.2076075E0,7.553571E0,2.01448E0,2.294999E0,3.3736875E0,8.374045E0,4.94766E0,1.1213868E0,1.0862207E0,1.7901583E0,5.763413E0,1.1139808E0,1.1810181E0,2.853631E0,2.094029E0,1.0056059E0,1.0088742E0,2.815665E0,5.5583806E0,1.2020774E0,4.5613356E0,1.4331759E0,1.3824891E0,1.6249641E0,1.2286669E0,1.4501098E0,4.1082706E0,3.4387102E0,1.1226254E0,2.438435E0,1.6698354E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"37","size_leaf_vector":"0"}},{"base_weights":[3.0229543E-3,1.0660087E-1,-1.6737056E-1,-1.09515995E-1,2.4333718E-1,1.20080836E-1,-3.1550086E-1,-3.3980307E-1,1.9091254E-1,-6.2303808E-2,3.2789278E-1,-1.9044252E-2,3.8538364E-1,1.752116E-2,-4.4257584E-1,-4.7115222E-1,2.0475555E-2,1.0060247E-2,1.3903984E-1,-5.229821E-1,-1.1900249E-2,-1.6212795E-2,4.720082E-1,-3.3555813E-2,1.0761522E-1,-2.3053579E-1,7.3587246E-2,-1.6822624E-1,-4.3213915E-2,5.692842E-1,1.5516852E-2,-1.7514041E-1,-4.2705115E-2,-1.21978335E-1,2.3959935E-2,4.998481E-2,1.8724892E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,true,false,true,true,true,true,false,true,true,false,false,true,true,false,false,false,true,false,false,true,false,false,false,false,false,false,false,false,false,false,false,false,false,false],"id":26,"left_children":[1,3,5,7,11,9,13,15,23,-1,17,25,21,-1,19,27,-1,-1,-1,31,-1,-1,29,-1,-1,33,-1,-1,-1,35,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[5.513363E-1,6.002609E-1,5.4106206E-1,6.329948E-1,4.620087E-1,3.911559E-1,4.0948057E-1,3.005848E-1,2.5368232E-1,0E0,1.1576924E-1,3.4935564E-1,3.2856786E-1,0E0,1.7825437E-1,3.397703E-2,0E0,0E0,0E0,3.3539295E-2,0E0,0E0,2.4665499E-1,0E0,0E0,2.196543E-1,0E0,0E0,0E0,1.4981747E-2,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,10,10,14,14,12,12,8,8,11,11,15,15,22,22,19,19,25,25,29,29],"right_children":[2,4,6,8,12,10,14,16,24,-1,18,26,22,-1,20,28,-1,-1,-1,32,-1,-1,30,-1,-1,34,-1,-1,-1,36,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[2.1343677E0,-1.3897004E0,-3.2199953E0,1.3853186E0,-3.3030245E-1,-3.793531E0,-5.36903E0,6.5933764E-1,-3.7884623E-1,-6.2303808E-2,2.4850893E-1,4.426071E-2,-7.840069E0,1.752116E-2,4.0739455E0,2.8838427E0,2.0475555E-2,1.0060247E-2,1.3903984E-1,3.9791207E0,-1.1900249E-2,-1.6212795E-2,6.116941E0,-3.3555813E-2,1.0761522E-1,-1.2552595E-1,7.3587246E-2,-1.6822624E-1,-4.3213915E-2,-9.1522855E-1,1.5516852E-2,-1.7514041E-1,-4.2705115E-2,-1.21978335E-1,2.3959935E-2,4.998481E-2,1.8724892E-1],"split_indices":[117,36,109,73,192,9,42,279,208,0,129,373,16,0,78,109,0,0,0,97,0,0,139,0,0,76,0,0,0,402,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[2.922802E1,1.8410889E1,1.081713E1,7.21166E0,1.1199228E1,3.743027E0,7.074103E0,4.008199E0,3.2034607E0,1.4526117E0,2.2904153E0,4.237622E0,6.961606E0,1.9314407E0,5.142662E0,2.915195E0,1.0930041E0,1.1155875E0,1.1748278E0,4.038669E0,1.1039928E0,1.2083013E0,5.7533045E0,1.2612538E0,1.9422069E0,2.4251509E0,1.8124709E0,1.7276591E0,1.1875359E0,4.384081E0,1.3692238E0,3.0220258E0,1.0166432E0,1.3497078E0,1.075443E0,1.0099255E0,3.3741555E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"37","size_leaf_vector":"0"}},{"base_weights":[1.1281905E-3,-1.9393319E-1,8.620405E-2,6.344474E-2,-3.3334E-1,2.596903E-1,-9.6829064E-2,-4.7882536E-1,-6.6078617E-3,-7.534398E-2,3.5856354E-1,1.5824585E-1,-2.3858061E-1,-1.7452633E-2,-1.7387673E-1,1.2136054E-3,4.534743E-1,5.4147135E-2,-4.283321E-1,-8.295217E-2,8.014123E-2,-4.7493096E-2,3.2494295E-1,-3.136531E-2,3.2116275E-2,7.360381E-2,-6.10635E-2,1.2679967E-1,2.603358E-2,5.7060087E-1,1.7237311E-2,-1.2505272E-2,-1.6564353E-1,1.8964922E-1,4.7516398E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,true,false,false,false,true,false,false,false,false,false,false,true,true,true,false,false,false,false,true,false,false,false,false,false,false,true,false,false,false,false,false],"id":27,"left_children":[1,3,5,-1,7,9,11,13,19,-1,15,21,17,-1,-1,23,29,25,31,-1,-1,-1,27,-1,-1,-1,-1,-1,-1,33,-1,-1,-1,-1,-1],"loss_changes":[4.9554896E-1,5.831375E-1,6.850977E-1,0E0,3.3612603E-1,6.426908E-1,4.2976642E-1,1.7502046E-1,3.112577E-1,0E0,3.243401E-1,2.8753424E-1,4.3244055E-1,0E0,0E0,4.4920225E-2,3.1816626E-1,2.3644765E-1,1.9635206E-1,0E0,0E0,0E0,3.9156556E-2,0E0,0E0,0E0,0E0,0E0,0E0,4.521346E-2,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,4,4,5,5,6,6,7,7,10,10,12,12,8,8,11,11,15,15,17,17,22,22,16,16,18,18,29,29],"right_children":[2,4,6,-1,8,10,12,14,20,-1,16,22,18,-1,-1,24,30,26,32,-1,-1,-1,28,-1,-1,-1,-1,-1,-1,34,-1,-1,-1,-1,-1],"split_conditions":[-4.538929E0,-7.211624E0,7.3667336E-1,6.344474E-2,2.4769115E0,-7.274233E0,-2.5220025E0,-6.027588E-1,3.0027345E-1,-7.534398E-2,-4.7421246E0,-5.43623E-1,-1.2528253E0,-1.7452633E-2,-1.7387673E-1,2.735904E-1,5.0826125E0,9.03629E-2,-2.8763223E0,-8.295217E-2,8.014123E-2,-4.7493096E-2,-1.6246916E-1,-3.136531E-2,3.2116275E-2,7.360381E-2,-6.10635E-2,1.2679967E-1,2.603358E-2,6.0223824E-1,1.7237311E-2,-1.2505272E-2,-1.6564353E-1,1.8964922E-1,4.7516398E-2],"split_indices":[47,61,64,0,144,129,107,459,482,0,5,83,72,0,0,291,25,421,131,0,0,0,460,0,0,0,0,0,0,186,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[2.785392E1,8.070469E0,1.9783451E1,1.9327444E0,6.1377244E0,9.941589E0,9.841862E0,3.9246712E0,2.2130535E0,1.3624235E0,8.579165E0,3.4740748E0,6.367787E0,1.06327E0,2.861401E0,2.0129533E0,6.566212E0,2.7854078E0,3.582379E0,1.1092743E0,1.1037792E0,1.2157289E0,2.2583458E0,1.0129327E0,1.0000206E0,1.6265057E0,1.158902E0,1.0520096E0,1.2063363E0,4.727228E0,1.838984E0,1.1931224E0,2.3892567E0,3.6487136E0,1.0785147E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"35","size_leaf_vector":"0"}},{"base_weights":[4.2176511E-4,2.0281589E-1,-7.6269984E-2,3.1101662E-1,-4.448088E-2,-1.5911692E-1,2.1542789E-1,4.1354915E-1,-2.1747423E-2,2.2848235E-1,-2.3662393E-1,-4.2720035E-2,1.2425532E-1,9.143652E-3,1.5494502E-1,-5.0876024E-3,1.1715562E-1,-3.2359293E-1,1.09270565E-1,8.600904E-3,-4.189666E-1,9.541813E-2,-5.357967E-2,8.664393E-2,-8.271586E-2,-5.8299594E-2,-5.2189E-1,-6.42978E-2,4.1656733E-2,-6.049955E-3,-1.8191786E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,false,true,true,false,false,true,true,false,false,false,false,false,false,true,true,true,true,false,false,false,false,false,true,false,false,false,false],"id":28,"left_children":[1,3,5,7,-1,9,11,13,-1,15,17,-1,-1,-1,-1,-1,-1,19,21,23,25,-1,-1,-1,-1,27,29,-1,-1,-1,-1],"loss_changes":[4.423858E-1,3.4360746E-1,5.3386205E-1,2.7806908E-1,0E0,5.550167E-1,4.1440624E-1,1.8643641E-1,0E0,1.3737085E-1,4.681707E-1,0E0,0E0,0E0,0E0,0E0,0E0,3.758545E-1,2.754807E-1,3.665009E-1,3.0918312E-1,0E0,0E0,0E0,0E0,1.2687089E-1,2.8420305E-1,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,6,6,7,7,9,9,10,10,17,17,18,18,19,19,20,20,25,25,26,26],"right_children":[2,4,6,8,-1,10,12,14,-1,16,18,-1,-1,-1,-1,-1,-1,20,22,24,26,-1,-1,-1,-1,28,30,-1,-1,-1,-1],"split_conditions":[-5.9482584E0,5.512516E-1,6.080864E0,5.3664774E-1,-4.448088E-2,-8.9349375E0,-1.076448E0,-5.6109447E0,-2.1747423E-2,1.4554846E-1,5.4691277E0,-4.2720035E-2,1.2425532E-1,9.143652E-3,1.5494502E-1,-5.0876024E-3,1.1715562E-1,-5.0629854E0,-3.326217E-2,-3.1847525E-1,-4.307416E0,9.541813E-2,-5.357967E-2,8.664393E-2,-8.271586E-2,9.444046E-2,-6.01981E0,-6.42978E-2,4.1656733E-2,-6.049955E-3,-1.8191786E-1],"split_indices":[63,425,133,314,0,61,35,93,0,346,6,0,0,0,0,0,0,1,394,147,86,0,0,0,0,73,164,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[2.6504066E1,6.8312993E0,1.9672768E1,5.309192E0,1.522107E0,1.5675255E1,3.9975128E0,4.1273856E0,1.1818064E0,2.2875776E0,1.3387677E1,1.528767E0,2.4687457E0,1.1486772E0,2.9787085E0,1.2657107E0,1.0218669E0,1.0749406E1,2.638271E0,2.6007195E0,8.148686E0,1.4683843E0,1.1698868E0,1.3018646E0,1.2988548E0,2.15689E0,5.9917965E0,1.1554158E0,1.0014741E0,1.0422482E0,4.9495482E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"31","size_leaf_vector":"0"}},{"base_weights":[4.597686E-3,1.775408E-1,-8.9131005E-2,3.4828568E-1,-9.5438205E-2,-2.273679E-1,8.957062E-2,4.94062E-1,-2.7859883E-2,-3.377196E-1,6.813908E-2,5.2814167E-2,-2.7490553E-1,3.2445875E-1,-1.0262986E-1,2.9950444E-2,1.7583695E-1,-4.6345025E-2,8.5880965E-2,-2.731054E-1,5.637143E-2,1.3143277E-1,3.3356112E-3,-4.524827E-1,2.132548E-2,1.0279363E-3,-1.23499215E-1,1.7252007E-2,-1.3250317E-1,-1.633696E-1,-1.1101817E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,true,true,false,true,false,true,true,false,true,true,false,false,false,false,false,true,false,false,false,true,false,false,false,false,false,false,false],"id":29,"left_children":[1,3,5,7,11,9,13,15,-1,23,17,-1,25,21,19,-1,-1,-1,-1,27,-1,-1,-1,29,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[4.416459E-1,4.7917068E-1,4.5766497E-1,4.38228E-1,2.6601276E-1,3.5770965E-1,4.1425684E-1,1.13652706E-1,0E0,3.906452E-1,2.2014971E-1,0E0,1.1862159E-1,1.3891983E-1,3.188054E-1,0E0,0E0,0E0,0E0,2.378012E-1,0E0,0E0,0E0,2.1329427E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,10,10,14,14,13,13,9,9,12,12,19,19,23,23],"right_children":[2,4,6,8,12,10,14,16,-1,24,18,-1,26,22,20,-1,-1,-1,-1,28,-1,-1,-1,30,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[-3.1927748E0,1.7179898E0,2.1221077E-1,5.0912046E0,-4.4371414E-1,5.9348536E0,-1.8185403E0,-5.481591E-1,-2.7859883E-2,3.7776232E0,2.010748E0,5.2814167E-2,-1.3522567E-1,4.3953192E-1,2.0455177E-1,2.9950444E-2,1.7583695E-1,-4.6345025E-2,8.5880965E-2,-1.6121107E-1,5.637143E-2,1.3143277E-1,3.3356112E-3,7.6623154E-1,2.132548E-2,1.0279363E-3,-1.23499215E-1,1.7252007E-2,-1.3250317E-1,-1.633696E-1,-1.1101817E-2],"split_indices":[106,132,76,102,159,122,64,397,0,156,4,0,233,285,439,0,0,0,0,325,0,0,0,484,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[2.5269497E1,8.567328E0,1.670217E1,5.100907E0,3.4664216E0,9.263748E0,7.4384217E0,3.7438388E0,1.3570677E0,6.640956E0,2.6227922E0,1.3871287E0,2.079293E0,3.0378003E0,4.4006214E0,1.1033746E0,2.6404643E0,1.4424376E0,1.1803546E0,2.8111355E0,1.5894859E0,1.9370344E0,1.100766E0,5.1018715E0,1.5390843E0,1.00627E0,1.073023E0,1.1717994E0,1.6393361E0,3.9219456E0,1.1799259E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"31","size_leaf_vector":"0"}},{"base_weights":[6.025361E-3,5.6411713E-2,-3.014259E-1,-1.3803196E-1,1.4069304E-1,-3.4642864E-3,-1.2729308E-1,-3.138987E-1,1.4447026E-1,-5.2275952E-2,2.8120396E-1,-4.743929E-3,-1.3850358E-1,-2.3300199E-1,1.736175E-1,-2.9386384E-2,8.536382E-2,3.688228E-1,-3.1238573E-2,1.4839927E-2,-1.16374485E-1,-4.344163E-2,1.0386219E-1,-2.7627487E-2,4.657962E-1,3.5083946E-4,1.6536708E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,true,false,false,true,true,false,true,false,false,true,false,false,false,false,false,false,false,false,false,false,true,false,false],"id":30,"left_children":[1,3,5,7,9,-1,-1,11,15,13,17,-1,-1,19,21,-1,-1,23,-1,-1,-1,-1,-1,-1,25,-1,-1],"loss_changes":[4.0628582E-1,3.8705122E-1,1.267575E-1,4.0744174E-1,4.5245683E-1,0E0,0E0,2.0445225E-1,1.4608362E-1,3.5259372E-1,3.5911512E-1,0E0,0E0,2.2373796E-1,2.683814E-1,0E0,0E0,3.9865375E-1,0E0,0E0,0E0,0E0,0E0,0E0,2.6691675E-1,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,7,7,9,9,8,8,10,10,13,13,14,14,17,17,24,24],"right_children":[2,4,6,8,10,-1,-1,12,16,14,18,-1,-1,20,22,-1,-1,24,-1,-1,-1,-1,-1,-1,26,-1,-1],"split_conditions":[7.678087E0,-4.469984E0,-8.0668785E-2,2.3767297E0,-8.801168E-1,-3.4642864E-3,-1.2729308E-1,-3.1029463E-1,-1.2453327E0,4.413488E-1,6.648756E0,-4.743929E-3,-1.3850358E-1,-6.268492E-1,-5.1698046E0,-2.9386384E-2,8.536382E-2,-6.4911337E0,-3.1238573E-2,1.4839927E-2,-1.16374485E-1,-4.344163E-2,1.0386219E-1,-2.7627487E-2,-7.508335E0,3.5083946E-4,1.6536708E-1],"split_indices":[27,154,226,150,11,0,0,40,153,158,69,0,0,150,55,0,0,103,0,0,0,0,0,0,135,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[2.412899E1,2.1432985E1,2.6960053E0,6.288097E0,1.5144888E1,1.1283213E0,1.567684E0,3.8069873E0,2.4811096E0,6.6458497E0,8.499039E0,1.6287177E0,2.1782696E0,3.6745543E0,2.9712954E0,1.0187292E0,1.4623804E0,6.9594173E0,1.5396211E0,1.5425498E0,2.1320045E0,1.100989E0,1.8703064E0,1.2184535E0,5.740964E0,1.0490466E0,4.6919174E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"27","size_leaf_vector":"0"}},{"base_weights":[-1.2781012E-3,9.626126E-2,-1.5273552E-1,2.072986E-1,-1.8833227E-1,1.7422834E-1,-2.941539E-1,2.9902428E-1,-8.5452326E-2,1.1059523E-1,-2.5970222E-2,-3.3304036E-1,4.2274687E-2,-1.5513118E-1,-5.458142E-2,3.8126162E-1,-2.5259169E-2,-1.2746453E-1,-2.4453865E-2,3.2063376E-2,-8.413951E-2,-8.904057E-2,1.4182124E-1,1.8787984E-2,4.8243132E-1,-2.9229766E-2,9.075489E-2,1.755536E-1,4.4549357E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,true,true,false,true,true,false,false,true,false,false,false,false,false,false,false,false,false,false,true,false,true,false,false,false,false],"id":31,"left_children":[1,3,5,7,11,9,13,15,19,-1,-1,17,-1,-1,21,23,-1,-1,-1,-1,-1,-1,25,-1,27,-1,-1,-1,-1],"loss_changes":[3.714413E-1,5.2258813E-1,5.065185E-1,3.3391076E-1,2.765109E-1,2.1449172E-1,3.5876226E-1,3.1466407E-1,1.6190009E-1,0E0,0E0,4.956132E-2,0E0,0E0,2.6142335E-1,2.1358097E-1,0E0,0E0,0E0,0E0,0E0,0E0,1.533769E-1,0E0,1.0493362E-1,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,11,11,8,8,14,14,15,15,22,22,24,24],"right_children":[2,4,6,8,12,10,14,16,20,-1,-1,18,-1,-1,22,24,-1,-1,-1,-1,-1,-1,26,-1,28,-1,-1,-1,-1],"split_conditions":[3.0724077E0,4.3753586E0,-4.2901797E0,5.986258E-1,4.2916093E0,6.663286E-2,-1.6451392E0,8.7597185E-1,9.002538E-1,1.1059523E-1,-2.5970222E-2,-1.2080811E-1,4.2274687E-2,-1.5513118E-1,-4.1138897E0,-4.8934617E0,-2.5259169E-2,-1.2746453E-1,-2.4453865E-2,3.2063376E-2,-8.413951E-2,-8.904057E-2,-1.5307587E-1,1.8787984E-2,1.6865804E0,-2.9229766E-2,9.075489E-2,1.755536E-1,4.4549357E-2],"split_indices":[25,109,27,425,55,325,94,314,80,0,0,461,0,0,31,47,0,0,0,0,0,0,477,0,29,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[2.3147926E1,1.4301876E1,8.846049E0,1.0483298E1,3.8185782E0,2.6008332E0,6.2452164E0,7.965951E0,2.5173473E0,1.2530943E0,1.3477389E0,2.6446896E0,1.1738886E0,2.6347919E0,3.6104245E0,6.5627365E0,1.4032145E0,1.4324472E0,1.2122424E0,1.4949272E0,1.0224202E0,1.3877463E0,2.2226782E0,1.9717417E0,4.590995E0,1.0512242E0,1.171454E0,2.9354203E0,1.6555746E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"29","size_leaf_vector":"0"}},{"base_weights":[-2.6474805E-3,9.311093E-2,-1.5534854E-1,1.5890145E-1,-2.139803E-1,-3.3930263E-1,2.039829E-2,2.1918572E-1,-5.3553704E-2,-1.109547E-2,-1.2790824E-1,-8.9879476E-2,-7.200083E-3,2.2100091E-1,-7.5788565E-2,3.0911678E-1,-5.5935487E-2,-1.1545783E-2,1.1313263E-1,4.1759324E-1,2.5592953E-2,4.720848E-2,-8.282274E-2,1.5108575E-1,2.6772678E-2,8.294653E-2,-6.356268E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,true,false,false,true,false,false,false,false,false,false,true,false,false,true,false,false,true,false,false,false,false,false,false,false],"id":32,"left_children":[1,3,5,7,11,9,13,15,-1,-1,-1,-1,-1,17,-1,19,21,-1,-1,23,25,-1,-1,-1,-1,-1,-1],"loss_changes":[3.5360736E-1,3.346265E-1,3.0802608E-1,3.000625E-1,4.244414E-2,1.0549134E-1,3.6836904E-1,2.9767936E-1,0E0,0E0,0E0,0E0,0E0,1.693699E-1,0E0,2.5670135E-1,2.1429312E-1,0E0,0E0,1.3083053E-1,2.6539037E-1,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,13,13,15,15,16,16,19,19,20,20],"right_children":[2,4,6,8,12,10,14,16,-1,-1,-1,-1,-1,18,-1,20,22,-1,-1,24,26,-1,-1,-1,-1,-1,-1],"split_conditions":[2.0757966E0,7.6879067E0,-1.2573159E-1,1.1948591E0,4.0504217E-1,-4.237048E-1,1.4399707E0,5.244933E0,-5.3553704E-2,-1.109547E-2,-1.2790824E-1,-8.9879476E-2,-7.200083E-3,1.9363973E-2,-7.5788565E-2,3.591281E0,1.0644693E0,-1.1545783E-2,1.1313263E-1,5.5386806E-1,-1.08851306E-1,4.720848E-2,-8.282274E-2,1.5108575E-1,2.6772678E-2,8.294653E-2,-6.356268E-2],"split_indices":[117,75,199,272,149,307,109,9,0,0,0,0,0,490,0,81,37,0,0,441,343,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[2.2193357E1,1.3879686E1,8.313671E0,1.1828199E1,2.0514874E0,3.60729E0,4.7063813E0,1.0332526E1,1.4956735E0,1.1251006E0,2.4821894E0,1.0164205E0,1.035067E0,2.8228698E0,1.8835114E0,7.6939707E0,2.6385555E0,1.3433617E0,1.479508E0,5.22284E0,2.4711308E0,1.4849459E0,1.1536096E0,3.7155945E0,1.5072454E0,1.121696E0,1.3494349E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"27","size_leaf_vector":"0"}},{"base_weights":[-5.911931E-3,-7.876446E-2,1.9813468E-1,1.678906E-1,-1.6153067E-1,3.548049E-1,-4.710592E-2,1.1755882E-1,-7.021607E-2,-2.6597765E-1,1.2114711E-1,1.3582051E-1,9.599184E-4,4.1185394E-2,-7.5799E-2,-2.7048385E-3,-3.679435E-1,2.9189062E-1,-6.3669495E-2,7.0615456E-2,-6.972878E-2,-4.6135062E-1,-1.7502524E-2,1.8433552E-2,1.1760001E-1,-4.710998E-2,-1.6834952E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,true,true,false,false,true,true,true,false,false,false,false,true,false,false,false,false,false,false,false,false,false,false,false],"id":33,"left_children":[1,3,5,7,9,11,-1,-1,13,15,17,-1,-1,-1,-1,19,21,23,-1,-1,-1,25,-1,-1,-1,-1,-1],"loss_changes":[3.4878406E-1,3.7890062E-1,3.9965612E-1,2.8455025E-1,4.2820883E-1,1.5680534E-1,0E0,0E0,1.5523465E-1,2.7030778E-1,3.0703488E-1,0E0,0E0,0E0,0E0,2.6412946E-1,1.8354172E-1,5.1313102E-2,0E0,0E0,0E0,7.730174E-2,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,5,5,8,8,9,9,10,10,15,15,16,16,17,17,21,21],"right_children":[2,4,6,8,10,12,-1,-1,14,16,18,-1,-1,-1,-1,20,22,24,-1,-1,-1,26,-1,-1,-1,-1,-1],"split_conditions":[5.923786E0,-5.676882E0,3.1069074E0,-8.9565706E-1,3.8640652E0,3.1755936E-1,-4.710592E-2,1.1755882E-1,-3.4477767E-2,-3.0461774E0,4.7948523E0,1.3582051E-1,9.599184E-4,4.1185394E-2,-7.5799E-2,-2.7404404E-1,5.077216E0,2.979125E-1,-6.3669495E-2,7.0615456E-2,-6.972878E-2,-3.1121278E-1,-1.7502524E-2,1.8433552E-2,1.1760001E-1,-4.710998E-2,-1.6834952E-1],"split_indices":[161,106,140,122,156,200,0,0,261,125,134,0,0,0,0,87,4,171,0,0,0,226,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[2.1413124E1,1.623175E1,5.1813745E0,3.8197784E0,1.2411972E1,3.5960317E0,1.5853429E0,1.6355624E0,2.184216E0,9.106338E0,3.3056343E0,2.5876799E0,1.0083519E0,1.1377592E0,1.0464568E0,2.828852E0,6.277486E0,2.2683313E0,1.037303E0,1.377034E0,1.451818E0,4.445986E0,1.8314999E0,1.1757076E0,1.0926237E0,1.7336454E0,2.7123404E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"27","size_leaf_vector":"0"}},{"base_weights":[-3.8824852E-3,1.1336239E-1,-1.2498749E-1,3.048837E-1,-3.976089E-2,-2.6970154E-1,7.725436E-2,4.156421E-1,-1.07221715E-2,-3.769672E-1,1.32605685E-2,-2.171456E-1,1.7760983E-1,-5.184753E-2,2.40663E-1,1.5669164E-1,2.7504731E-2,1.5451406E-2,-1.02877274E-1,9.6455944E-4,1.0532318E-1,-5.363132E-3,-1.4152242E-1,1.2088257E-1,-2.3922069E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,true,true,true,true,true,false,false,false,true,true,false,true,false,false,false,false,false,false,false,false,false,false],"id":34,"left_children":[1,3,5,7,11,9,13,15,-1,21,-1,17,23,-1,19,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[3.2064217E-1,3.4912014E-1,3.4549567E-1,2.1182305E-1,3.1890166E-1,2.4188101E-1,2.6243463E-1,9.098792E-2,0E0,1.6479874E-1,0E0,1.6814306E-1,2.5347915E-1,0E0,9.552069E-2,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,11,11,14,14,9,9,12,12],"right_children":[2,4,6,8,12,10,14,16,-1,22,-1,18,24,-1,20,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[3.3998775E-1,-2.6720266E0,3.497243E-1,4.2825003E0,2.0372262E0,2.328743E0,-2.8441608E-1,3.0684566E-1,-1.07221715E-2,-3.4646683E0,1.32605685E-2,-1.6891351E0,-8.003197E-1,-5.184753E-2,-2.2960193E-1,1.5669164E-1,2.7504731E-2,1.5451406E-2,-1.02877274E-1,9.6455944E-4,1.0532318E-1,-5.363132E-3,-1.4152242E-1,1.2088257E-1,-2.3922069E-2],"split_indices":[159,63,129,134,137,73,446,157,0,69,0,61,134,0,394,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[2.0584236E1,1.0491285E1,1.0092951E1,4.220869E0,6.2704163E0,5.6887846E0,4.404166E0,3.0189738E0,1.2018951E0,4.090199E0,1.5985855E0,3.4533534E0,2.8170629E0,1.7177336E0,2.6864326E0,1.8105859E0,1.208388E0,1.2895402E0,2.163813E0,1.1793464E0,1.5070862E0,1.1023047E0,2.9878943E0,1.2003334E0,1.6167295E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"25","size_leaf_vector":"0"}},{"base_weights":[2.3357475E-3,-7.440994E-2,1.7995548E-1,-2.1564817E-1,6.9920294E-2,-7.016039E-2,3.0832002E-1,3.2581102E-2,-3.1384388E-1,-6.435411E-2,3.2664813E-2,-9.005836E-2,1.8729912E-1,1.3379413E-1,1.9853722E-2,-1.3690016E-1,-4.6374664E-2,-9.3608275E-2,3.7580308E-1,-7.555194E-2,5.2416276E-2,4.0510498E-2,-9.1655254E-2,1.4891742E-1,1.553332E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,true,true,true,true,false,true,false,false,false,true,false,false,false,true,true,true,false,false,false,false,false,false],"id":35,"left_children":[1,3,5,7,11,9,13,-1,15,-1,-1,-1,17,-1,-1,-1,19,21,23,-1,-1,-1,-1,-1,-1],"loss_changes":[2.9791346E-1,3.2660833E-1,2.3321149E-1,2.8781098E-1,4.1739917E-1,1.0217738E-1,1.2182975E-1,0E0,2.2174418E-1,0E0,0E0,0E0,4.0201187E-1,0E0,0E0,0E0,1.9520146E-1,2.14836E-1,1.4407653E-1,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,8,8,12,12,16,16,17,17,18,18],"right_children":[2,4,6,8,12,10,14,-1,16,-1,-1,-1,18,-1,-1,-1,20,22,24,-1,-1,-1,-1,-1,-1],"split_conditions":[4.8019023E0,-7.7319765E-1,-2.564472E0,-6.235343E-1,-7.0387616E0,-2.1069776E-1,-1.324281E0,3.2581102E-2,3.155585E-1,-6.435411E-2,3.2664813E-2,-9.005836E-2,-1.6430502E0,1.3379413E-1,1.9853722E-2,-1.3690016E-1,7.05818E-2,1.5607958E-1,3.7001532E-1,-7.555194E-2,5.2416276E-2,4.0510498E-2,-9.1655254E-2,1.4891742E-1,1.553332E-2],"split_indices":[70,94,28,403,3,264,80,0,231,0,0,0,38,0,0,0,385,258,173,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.987239E1,1.428236E1,5.5900307E0,6.9687605E0,7.3135996E0,2.049686E0,3.5403447E0,1.5952142E0,5.373546E0,1.0250936E0,1.0245924E0,1.3859586E0,5.927641E0,1.7204336E0,1.8199111E0,3.0451658E0,2.3283803E0,2.5825527E0,3.3450885E0,1.1347749E0,1.1936054E0,1.4167233E0,1.1658294E0,2.050156E0,1.2949322E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"25","size_leaf_vector":"0"}},{"base_weights":[2.7172435E-3,-9.726029E-2,1.4231163E-1,-2.1912967E-1,1.2139289E-1,2.7603924E-1,-7.913414E-2,-3.6463204E-1,1.3794469E-2,5.0898003E-3,1.3303699E-1,-4.2171903E-2,2.2652407E-1,-8.0046594E-2,3.8898755E-2,-1.4237635E-1,-9.600708E-3,-5.1537607E-2,5.2796103E-2,5.8465343E-2,-5.851633E-2,1.0356701E-1,-1.6379999E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,true,true,false,false,true,true,true,false,false,true,false,false,false,false,false,false,false,false,false,false],"id":36,"left_children":[1,3,5,7,11,9,13,15,19,17,-1,-1,21,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[2.9468355E-1,3.548755E-1,2.8254992E-1,2.8520226E-1,1.7099205E-1,2.5683752E-1,1.9495207E-1,1.7224789E-1,1.9466913E-1,1.2675299E-1,0E0,0E0,1.5194052E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,9,9,8,8,12,12],"right_children":[2,4,6,8,12,10,14,16,20,18,-1,-1,22,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[1.7952664E0,2.5310683E0,2.4966896E-1,4.6205103E-1,-4.992193E-1,8.5606337E-1,4.0006638E-3,3.1579208E0,1.0522923E0,-1.1259531E-1,1.3303699E-1,-4.2171903E-2,2.6623353E-1,-8.0046594E-2,3.8898755E-2,-1.4237635E-1,-9.600708E-3,-5.1537607E-2,5.2796103E-2,5.8465343E-2,-5.851633E-2,1.0356701E-1,-1.6379999E-2],"split_indices":[133,56,291,142,488,88,47,3,85,346,0,0,315,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.9122969E1,1.1319328E1,7.8036404E0,7.266863E0,4.0524654E0,4.7117543E0,3.091886E0,4.1247635E0,3.1420991E0,2.1929963E0,2.518758E0,1.0640202E0,2.9884453E0,1.4818993E0,1.6099868E0,2.7792616E0,1.3455018E0,1.0750566E0,1.1179397E0,1.718701E0,1.4233981E0,1.9409187E0,1.0475266E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"23","size_leaf_vector":"0"}},{"base_weights":[1.2021937E-3,-6.893868E-2,1.9287415E-1,-2.0670432E-1,5.6701027E-2,-3.368781E-2,2.9963803E-1,2.2252275E-2,-3.2350826E-1,2.275019E-1,-1.384942E-1,-8.468131E-3,1.2432266E-1,9.798954E-4,-4.0774414E-1,-2.3817621E-2,3.5350823E-1,-2.8694633E-1,3.4186848E-2,-1.4453824E-1,-3.0148204E-2,1.2798557E-1,2.8894616E-2,-1.1395328E-1,-1.3072535E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,true,true,false,false,false,true,false,false,false,false,false,false,false,true,true,false,false,false,false,false,false,false],"id":37,"left_children":[1,3,5,7,9,-1,11,-1,13,15,17,-1,-1,-1,19,-1,21,23,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[2.7542734E-1,2.7154374E-1,2.1241471E-1,2.6596147E-1,3.1942013E-1,0E0,1.7491153E-1,0E0,1.498512E-1,2.2080323E-1,2.0702961E-1,0E0,0E0,0E0,4.9363434E-2,0E0,2.8797537E-2,5.609989E-2,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,6,6,8,8,9,9,10,10,14,14,16,16,17,17],"right_children":[2,4,6,8,10,-1,12,-1,14,16,18,-1,-1,-1,20,-1,22,24,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[5.0261116E0,-3.4882784E-1,-5.318883E0,-5.8130485E-1,9.2807627E-1,-3.368781E-2,-4.437579E0,2.2252275E-2,-5.511676E0,-4.415324E0,1.7863046E-1,-8.468131E-3,1.2432266E-1,9.798954E-4,4.017255E0,-2.3817621E-2,3.4605944E-1,-8.423879E-1,3.4186848E-2,-1.4453824E-1,-3.0148204E-2,1.2798557E-1,2.8894616E-2,-1.1395328E-1,-1.3072535E-2],"split_indices":[5,11,160,265,132,0,66,0,81,156,222,0,0,0,114,0,368,144,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.8497929E1,1.401103E1,4.4868984E0,6.375258E0,7.635772E0,1.1494892E0,3.3374093E0,1.9796824E0,4.3955755E0,3.984085E0,3.6516871E0,1.060874E0,2.2765353E0,1.0978668E0,3.297709E0,1.2673453E0,2.7167397E0,2.2134404E0,1.4382467E0,2.199521E0,1.0981878E0,1.6024687E0,1.1142709E0,1.1961092E0,1.0173312E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"25","size_leaf_vector":"0"}},{"base_weights":[-7.6027826E-4,-1.1096478E-1,1.2129531E-1,7.621898E-2,-2.6048994E-1,2.0212077E-1,-6.37553E-2,1.9132015E-1,-6.012501E-2,9.820559E-3,-3.656911E-1,3.2954544E-1,-1.3956114E-2,-1.38361575E-2,9.44789E-2,4.3589908E-1,-1.2086446E-2,-1.6690128E-2,-1.3327524E-1,5.1796805E-2,-8.388439E-2,3.3369526E-2,1.580771E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,true,true,false,true,false,false,true,false,false,false,false,true,false,false,false,false,false,false,false],"id":38,"left_children":[1,3,5,7,9,11,-1,13,-1,-1,17,15,19,-1,-1,21,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[2.6810282E-1,3.1291693E-1,2.9819918E-1,2.1103299E-1,1.985153E-1,2.2841245E-1,0E0,1.4194867E-1,0E0,0E0,8.5692406E-2,2.2028291E-1,2.4532816E-1,0E0,0E0,6.218481E-2,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,5,5,7,7,11,11,10,10,12,12,15,15],"right_children":[2,4,6,8,10,12,-1,14,-1,-1,18,16,20,-1,-1,22,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[2.0019817E-1,-2.0152843E-1,7.476996E0,5.879962E0,-2.5611613E0,2.5899909E0,-6.37553E-2,9.090599E-1,-6.012501E-2,9.820559E-3,-4.652751E0,5.152213E0,4.2942435E-1,-1.38361575E-2,9.44789E-2,-2.6141986E-1,-1.2086446E-2,-1.6690128E-2,-1.3327524E-1,5.1796805E-2,-8.388439E-2,3.3369526E-2,1.580771E-1],"split_indices":[19,80,147,53,134,132,0,127,0,0,12,134,226,0,0,453,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.7931072E1,9.470755E0,8.460318E0,4.4234715E0,5.047283E0,7.1287556E0,1.3315624E0,3.341543E0,1.0819285E0,1.5145737E0,3.5327096E0,4.1539574E0,2.974798E0,1.358635E0,1.982908E0,3.0874567E0,1.0665007E0,1.0594542E0,2.4732554E0,1.9529964E0,1.0218016E0,1.1626177E0,1.924839E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"23","size_leaf_vector":"0"}},{"base_weights":[-2.1931538E-3,-6.300898E-2,2.0495531E-1,-1.3527167E-1,1.9957851E-1,9.5813185E-2,-7.3550437E-3,7.052273E-2,-2.3072225E-1,-1.3591459E-2,1.08976796E-1,2.140855E-1,-4.0985413E-2,-3.1049475E-1,9.866441E-3,4.176149E-3,9.6364334E-2,-4.3228114E-1,-3.7546173E-2,-3.2524463E-2,-1.4906408E-1,-5.2694958E-2,3.8338535E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,true,false,false,true,true,false,false,true,false,true,false,false,false,false,true,false,false,false,false],"id":39,"left_children":[1,3,5,7,9,-1,-1,11,13,-1,-1,15,-1,17,-1,-1,-1,19,21,-1,-1,-1,-1],"loss_changes":[2.4415554E-1,3.1076375E-1,1.23386264E-1,2.5833964E-1,1.6200483E-1,0E0,0E0,1.6813998E-1,1.9216812E-1,0E0,0E0,6.303677E-2,0E0,2.0776159E-1,0E0,0E0,0E0,4.1621923E-2,9.492486E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,7,7,8,8,11,11,13,13,17,17,18,18],"right_children":[2,4,6,8,10,-1,-1,12,14,-1,-1,16,-1,18,-1,-1,-1,20,22,-1,-1,-1,-1],"split_conditions":[6.313111E0,7.6190567E0,2.9340289E0,-4.290621E-1,-1.1873546E0,9.5813185E-2,-7.3550437E-3,1.5410209E-1,4.4004354E0,-1.3591459E-2,1.08976796E-1,4.184513E-1,-4.0985413E-2,3.4653804E-1,9.866441E-3,4.176149E-3,9.6364334E-2,-4.891331E-1,1.5897489E0,-3.2524463E-2,-1.4906408E-1,-5.2694958E-2,3.8338535E-2],"split_indices":[154,161,12,280,108,0,0,210,73,0,0,42,0,304,0,0,0,497,140,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.7355556E1,1.3954533E1,3.4010239E0,1.1323277E1,2.6312551E0,2.0079873E0,1.3930366E0,3.6705709E0,7.6527066E0,1.343863E0,1.2873921E0,2.1481965E0,1.5223744E0,5.7383423E0,1.9143645E0,1.1428257E0,1.0053707E0,3.5642643E0,2.174078E0,1.0380905E0,2.5261738E0,1.1506412E0,1.0234368E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"23","size_leaf_vector":"0"}},{"base_weights":[1.2471467E-3,7.076734E-2,-1.8830353E-1,-6.2175732E-2,1.9855213E-1,-3.0521876E-1,3.7708234E-2,-2.3279986E-1,1.3777804E-1,-3.8093014E-3,3.5416418E-1,-2.2613485E-3,-1.214631E-1,7.4528484E-3,-1.00583896E-1,5.6134503E-2,-5.72698E-2,-2.2079146E-2,1.00071944E-1,3.2423794E-2,1.2878715E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,true,true,false,true,true,true,true,false,false,false,false,false,false,false,false,false,false],"id":40,"left_children":[1,3,5,7,9,11,-1,13,17,15,19,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[2.489146E-1,2.4690606E-1,2.2451197E-1,2.8892845E-1,2.3055562E-1,1.1521441E-1,0E0,1.2540185E-1,1.9400972E-1,1.840813E-1,2.9868662E-2,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,5,5,7,7,9,9,8,8,10,10],"right_children":[2,4,6,8,10,12,-1,14,18,16,20,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[6.333127E0,3.3696353E-1,4.78869E0,8.1120896E-1,2.9762983E-2,-1.7632524E0,3.7708234E-2,-3.489713E-1,1.9111507E-1,7.441501E-2,-7.258382E-1,-2.2613485E-3,-1.214631E-1,7.4528484E-3,-1.00583896E-1,5.6134503E-2,-5.72698E-2,-2.2079146E-2,1.00071944E-1,3.2423794E-2,1.2878715E-1],"split_indices":[110,78,31,36,161,132,0,390,395,275,131,0,0,0,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.6877975E1,1.2807364E1,4.0706115E0,6.5286283E0,6.278736E0,2.9865465E0,1.084065E0,3.4340358E0,3.0945923E0,3.1534452E0,3.1252909E0,1.0188525E0,1.967694E0,1.1928059E0,2.24123E0,1.5606636E0,1.5927817E0,1.7882049E0,1.3063874E0,1.3013157E0,1.8239752E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"21","size_leaf_vector":"0"}},{"base_weights":[4.2696074E-3,1.196751E-1,-1.1572093E-1,2.0179597E-1,-4.766409E-2,-2.6906613E-1,4.4953596E-2,2.9072553E-1,-4.1113734E-2,1.7974725E-2,-1.22566186E-1,-1.219026E-1,7.216233E-2,-4.0417593E-3,3.7067023E-1,-8.221304E-2,2.6998637E-2,2.6974078E-2,1.3459674E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,true,false,false,true,true,false,false,false,true,false,false,true,false,false,false,false],"id":41,"left_children":[1,3,5,7,-1,9,11,13,-1,-1,-1,15,-1,-1,17,-1,-1,-1,-1],"loss_changes":[2.547177E-1,2.461312E-1,2.3473412E-1,2.7090126E-1,0E0,2.4321696E-1,2.0129152E-1,1.6049516E-1,0E0,0E0,0E0,1.3773078E-1,0E0,0E0,7.1293E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,6,6,7,7,11,11,14,14],"right_children":[2,4,6,8,-1,10,12,14,-1,-1,-1,16,-1,-1,18,-1,-1,-1,-1],"split_conditions":[3.3998775E-1,6.1989408E0,-5.077629E-1,8.021124E-1,-4.766409E-2,-2.4967759E0,6.3025975E-1,-5.809846E0,-4.1113734E-2,1.7974725E-2,-1.22566186E-1,-2.13706E0,7.216233E-2,-4.0417593E-3,-3.443604E-1,-8.221304E-2,2.6998637E-2,2.6974078E-2,1.3459674E-1],"split_indices":[159,164,47,349,0,69,97,160,0,0,0,142,0,0,315,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.6394304E1,8.358156E0,8.036148E0,6.6679435E0,1.690213E0,3.766683E0,4.269465E0,5.3942227E0,1.2737209E0,1.2913932E0,2.47529E0,2.507255E0,1.7622099E0,1.2956425E0,4.0985804E0,1.2887005E0,1.2185546E0,1.3589958E0,2.7395844E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"19","size_leaf_vector":"0"}},{"base_weights":[2.125658E-3,-1.465539E-1,8.069322E-2,-2.6470053E-1,4.187336E-2,-6.769218E-2,2.0009018E-1,1.5054314E-2,-3.6360756E-1,4.6489038E-2,-2.0848449E-1,3.1280577E-1,-4.0765297E-2,-1.3510461E-1,-2.595401E-2,-1.0349802E-1,1.0260916E-2,2.557679E-3,1.1661048E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,false,true,true,false,true,false,false,true,false,false,false,false,false,false,false],"id":42,"left_children":[1,3,5,7,-1,9,11,-1,13,-1,15,17,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[2.0961349E-1,2.468462E-1,2.2123478E-1,1.6586128E-1,0E0,2.1958466E-1,2.9807612E-1,0E0,5.064532E-2,0E0,1.471724E-1,1.2038022E-1,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,6,6,8,8,10,10,11,11],"right_children":[2,4,6,8,-1,10,12,-1,14,-1,16,18,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[-3.208375E0,5.5464067E0,-2.0247204E0,-3.900105E0,4.187336E-2,-2.0728579E0,5.2633305E0,1.5054314E-2,2.3990084E-1,4.6489038E-2,1.5137277E0,-4.5888276E0,-4.0765297E-2,-1.3510461E-1,-2.595401E-2,-1.0349802E-1,1.0260916E-2,2.557679E-3,1.1661048E-1],"split_indices":[153,70,55,1,0,164,141,0,388,0,5,35,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.5931851E1,5.2090445E0,1.0722807E1,3.7397602E0,1.4692842E0,4.974097E0,5.74871E0,1.0116625E0,2.7280977E0,1.8878726E0,3.0862243E0,4.356211E0,1.3924991E0,1.6015091E0,1.1265886E0,1.7054018E0,1.3808225E0,1.0917038E0,3.2645075E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"19","size_leaf_vector":"0"}},{"base_weights":[-3.090864E-3,-1.0122528E-1,1.1749984E-1,-2.182287E-1,7.9297535E-2,-1.326312E-1,2.314401E-1,-2.9991955E-1,2.627715E-2,-7.518453E-2,1.5071931E-2,-3.2425012E-2,6.909446E-2,-1.6424851E-2,3.1288174E-1,1.66231E-2,-1.20986335E-1,5.76102E-3,1.1782491E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,true,false,true,true,false,false,false,false,false,false,true,false,false,false,false],"id":43,"left_children":[1,3,5,7,11,9,13,15,-1,-1,-1,-1,-1,-1,17,-1,-1,-1,-1],"loss_changes":[2.0718952E-1,2.2082938E-1,2.5490636E-1,1.7898837E-1,1.528777E-1,7.942791E-2,1.5159836E-1,2.0784435E-1,0E0,0E0,0E0,0E0,0E0,0E0,1.0129395E-1,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,14,14],"right_children":[2,4,6,8,12,10,14,16,-1,-1,-1,-1,-1,-1,18,-1,-1,-1,-1],"split_conditions":[3.658762E-1,1.4676833E0,-4.2306933E0,4.732904E0,-1.7467172E-1,-2.1124634E-1,-6.161338E0,-6.0904665E0,2.627715E-2,-7.518453E-2,1.5071931E-2,-3.2425012E-2,6.909446E-2,-1.6424851E-2,-5.62143E-1,1.66231E-2,-1.20986335E-1,5.76102E-3,1.1782491E-1],"split_indices":[76,154,73,6,362,400,78,121,0,0,0,0,0,0,374,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.5502781E1,8.6357565E0,6.8670254E0,5.1129785E0,3.5227783E0,2.0977747E0,4.7692504E0,4.050338E0,1.0626407E0,1.0499349E0,1.0478399E0,1.6989877E0,1.8237906E0,1.1291418E0,3.6401086E0,1.0173011E0,3.0330367E0,1.0435106E0,2.5965981E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"19","size_leaf_vector":"0"}},{"base_weights":[-1.9793482E-3,-9.3255155E-2,1.24616094E-1,6.3200615E-2,-1.5804708E-1,-7.329727E-2,2.6675448E-1,4.3097377E-2,-2.6145807E-1,3.481761E-2,-7.432043E-2,-2.401791E-3,1.2067054E-1,-3.41098E-1,1.4308008E-2,-2.416605E-2,-1.23609856E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,true,true,true,false,false,false,false,false,false,false,false,false,false],"id":44,"left_children":[1,3,5,-1,7,9,11,-1,13,-1,-1,-1,-1,15,-1,-1,-1],"loss_changes":[1.9773746E-1,2.2885415E-1,2.2141817E-1,0E0,3.1078243E-1,1.5373722E-1,1.6671082E-1,0E0,1.8557143E-1,0E0,0E0,0E0,0E0,7.013196E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,4,4,5,5,6,6,8,8,13,13],"right_children":[2,4,6,-1,8,10,12,-1,14,-1,-1,-1,-1,16,-1,-1,-1],"split_conditions":[1.7952664E0,-9.59462E0,-1.77696E-1,6.3200615E-2,-5.9482584E0,2.1450875E-2,-1.9703188E0,4.3097377E-2,5.150621E0,3.481761E-2,-7.432043E-2,-2.401791E-3,1.2067054E-1,-3.3911133E0,1.4308008E-2,-2.416605E-2,-1.23609856E-1],"split_indices":[133,61,264,0,63,464,84,0,101,0,0,0,0,18,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.5106857E1,8.930961E0,6.175897E0,1.1737423E0,7.7572184E0,2.783909E0,3.391988E0,1.8807819E0,5.8764367E0,1.4953454E0,1.2885637E0,1.430923E0,1.961065E0,4.59054E0,1.2858968E0,1.4393586E0,3.1511815E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"17","size_leaf_vector":"0"}},{"base_weights":[1.616831E-3,8.9471966E-2,-1.3407286E-1,2.6108453E-1,-4.617188E-2,3.773165E-2,-2.457066E-1,-8.9845E-3,1.0979907E-1,7.576405E-2,-9.6132584E-2,1.6325727E-2,-3.3716512E-1,-4.0898856E-2,2.3250604E-1,-1.2710191E-1,-3.1822488E-2,9.946345E-2,5.08485E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,true,false,false,false,false,false,false,false,true,false,true,false,false,false,false],"id":45,"left_children":[1,3,5,7,9,-1,11,-1,-1,13,-1,-1,15,-1,17,-1,-1,-1,-1],"loss_changes":[1.9931863E-1,2.4804428E-1,2.2165167E-1,1.521998E-1,2.4206166E-1,0E0,1.5389976E-1,0E0,0E0,2.0844425E-1,0E0,0E0,3.3501655E-2,0E0,6.7539304E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,6,6,9,9,12,12,14,14],"right_children":[2,4,6,8,10,-1,12,-1,-1,14,-1,-1,16,-1,18,-1,-1,-1,-1],"split_conditions":[3.0724077E0,-2.2209193E-1,-4.2901797E0,-4.682658E0,6.648756E0,3.773165E-2,-6.959065E-1,-8.9845E-3,1.0979907E-1,-1.2349638E-2,-9.6132584E-2,1.6325727E-2,-7.332677E-2,-4.0898856E-2,4.1508533E-2,-1.2710191E-1,-3.1822488E-2,9.946345E-2,5.08485E-3],"split_indices":[25,274,27,160,69,0,305,0,0,264,0,0,439,0,296,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.471364E1,9.137795E0,5.5758443E0,3.625783E0,5.5120125E0,1.6375471E0,3.9382973E0,1.1500417E0,2.4757414E0,4.3166585E0,1.195354E0,1.0144182E0,2.923879E0,1.8897692E0,2.4268892E0,1.5210948E0,1.4027841E0,1.2941818E0,1.1327074E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"19","size_leaf_vector":"0"}},{"base_weights":[-1.2716847E-3,1.268724E-1,-9.29643E-2,-1.1786906E-1,2.995166E-1,5.461922E-2,-2.367742E-1,3.6577564E-2,-8.0865465E-2,-9.267915E-2,8.1314385E-2,1.2965932E-1,1.9309143E-2,-1.260041E-1,-6.6546895E-2,-7.8632176E-2,2.3573862E-2,3.0514281E-2,-6.0087774E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,true,false,true,true,false,false,true,false,false,false,false,true,false,false,false,false],"id":46,"left_children":[1,3,5,7,11,9,13,-1,-1,15,-1,-1,-1,-1,17,-1,-1,-1,-1],"loss_changes":[1.9179375E-1,3.2315147E-1,2.1470453E-1,1.6145864E-1,1.0562596E-1,2.0059864E-1,1.2808603E-1,0E0,0E0,1.3725474E-1,0E0,0E0,0E0,0E0,1.0547328E-1,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,9,9,14,14],"right_children":[2,4,6,8,12,10,14,-1,-1,16,-1,-1,-1,-1,18,-1,-1,-1,-1],"split_conditions":[-1.248637E0,-1.1932282E0,1.2084732E0,-1.7132452E-1,7.3667336E-1,1.8326879E0,-5.701872E0,3.6577564E-2,-8.0865465E-2,-1.5822412E0,8.1314385E-2,1.2965932E-1,1.9309143E-2,-1.260041E-1,-4.645598E-1,-7.8632176E-2,2.3573862E-2,3.0514281E-2,-6.0087774E-2],"split_indices":[86,101,134,395,64,112,92,0,0,19,0,0,0,0,122,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.4319251E1,5.8124495E0,8.506802E0,2.5354507E0,3.2769988E0,4.5044017E0,4.0024E0,1.0584046E0,1.4770461E0,3.020483E0,1.4839188E0,1.5592643E0,1.7177345E0,1.2208545E0,2.7815454E0,1.2516932E0,1.7687898E0,1.3378873E0,1.4436581E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"19","size_leaf_vector":"0"}},{"base_weights":[-2.5943723E-3,1.0172152E-1,-1.1046046E-1,2.1180059E-1,-8.305056E-2,1.04439124E-1,-2.0583837E-1,3.27206E-1,-1.586646E-2,6.365373E-2,-1.7373085E-2,4.571409E-2,-7.64358E-2,8.636627E-3,-3.1234786E-1,1.1829389E-1,2.5927301E-2,-1.253168E-1,-6.790474E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,true,false,true,true,false,false,false,false,false,false,false,false,false,false,false],"id":47,"left_children":[1,3,5,7,11,9,13,15,-1,-1,-1,-1,-1,-1,17,-1,-1,-1,-1],"loss_changes":[1.7902556E-1,1.8236534E-1,1.8158764E-1,1.809858E-1,1.9015354E-1,6.588704E-2,1.5259008E-1,2.8463572E-2,0E0,0E0,0E0,0E0,0E0,0E0,1.13773674E-1,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,7,7,14,14],"right_children":[2,4,6,8,12,10,14,16,-1,-1,-1,-1,-1,-1,18,-1,-1,-1,-1],"split_conditions":[3.3998775E-1,2.4939175E0,-4.342052E0,2.761402E-1,-1.8582222E0,-9.47557E-1,-1.387403E0,2.7036744E-1,-1.586646E-2,6.365373E-2,-1.7373085E-2,4.571409E-2,-7.64358E-2,8.636627E-3,1.8534964E0,1.1829389E-1,2.5927301E-2,-1.253168E-1,-6.790474E-3],"split_indices":[159,23,22,359,85,7,148,419,0,0,0,0,0,0,151,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.3911762E1,7.101225E0,6.8105373E0,4.358401E0,2.742824E0,2.064325E0,4.7462125E0,2.8706114E0,1.4877894E0,1.0563625E0,1.0079626E0,1.2044101E0,1.538414E0,1.7096888E0,3.0365236E0,1.7462784E0,1.124333E0,1.902642E0,1.1338816E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"19","size_leaf_vector":"0"}},{"base_weights":[-6.12477E-4,8.174083E-2,-1.3626572E-1,2.0769867E-1,-6.508475E-2,-2.6257172E-1,2.6513582E-2,-1.0121326E-2,3.0979016E-1,3.970413E-2,-1.8502368E-1,3.6260206E-3,-1.078198E-1,1.12810664E-1,2.5147468E-2,4.1040312E-3,-8.731973E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,false,false,false,false,true,false,false,false,false,false,false,false,false],"id":48,"left_children":[1,3,5,7,9,11,-1,-1,13,-1,15,-1,-1,-1,-1,-1,-1],"loss_changes":[1.7405783E-1,1.928997E-1,1.9015646E-1,1.4501746E-1,1.485734E-1,1.1120498E-1,0E0,0E0,2.738449E-2,0E0,8.145651E-2,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,5,5,8,8,10,10],"right_children":[2,4,6,8,10,12,-1,-1,14,-1,16,-1,-1,-1,-1,-1,-1],"split_conditions":[2.2912774E0,3.7443593E-2,2.3420334E0,-2.7694519E0,-3.947982E0,-3.1927748E0,2.6513582E-2,-1.0121326E-2,3.0490473E-1,3.970413E-2,7.479191E-3,3.6260206E-3,-1.078198E-1,1.12810664E-1,2.5147468E-2,4.1040312E-3,-8.731973E-2],"split_indices":[81,258,45,153,162,106,0,0,221,0,132,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.3583492E1,8.699542E0,4.8839498E0,4.4593697E0,4.2401724E0,3.0181599E0,1.8657898E0,1.5242351E0,2.9351344E0,1.5633246E0,2.6768477E0,1.0147947E0,2.0033653E0,1.7561586E0,1.1789758E0,1.2345383E0,1.4423094E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"17","size_leaf_vector":"0"}},{"base_weights":[-9.805965E-5,-8.5095845E-2,1.318953E-1,7.6832615E-2,-2.1202232E-1,-3.2937843E-2,2.2141315E-1,-4.660652E-2,1.8483932E-1,-3.1266296E-1,4.1628783E-3,3.0606845E-1,-5.9871897E-3,-1.0250689E-2,9.099589E-2,-6.2592346E-3,-1.2533838E-1,2.8197804E-2,1.11986615E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,true,true,false,true,false,true,false,false,false,false,false,false,false,false,false,false],"id":49,"left_children":[1,3,5,7,9,-1,11,-1,13,15,-1,17,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[1.7172466E-1,2.0763904E-1,1.5403484E-1,1.4971489E-1,1.290393E-1,0E0,1.02285445E-1,0E0,1.0766357E-1,1.0899314E-1,0E0,1.5463412E-2,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,6,6,8,8,9,9,11,11],"right_children":[2,4,6,8,10,-1,12,-1,14,16,-1,18,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[3.959839E-1,-8.5044146E-1,-5.1314635E0,-4.1108203E0,2.123695E0,-3.2937843E-2,1.812803E0,-4.660652E-2,-4.624412E-1,-2.0603967E0,4.1628783E-3,-7.0957375E-1,-5.9871897E-3,-1.0250689E-2,9.099589E-2,-6.2592346E-3,-1.2533838E-1,2.8197804E-2,1.11986615E-1],"split_indices":[177,117,47,36,156,0,109,0,137,139,0,100,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.3305986E1,8.310014E0,4.995972E0,3.8249464E0,4.4850674E0,1.2890882E0,3.706884E0,1.0751839E0,2.7497625E0,2.8370435E0,1.6480238E0,2.545916E0,1.160968E0,1.2151642E0,1.5345984E0,1.0688487E0,1.7681948E0,1.1899562E0,1.3559599E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"19","size_leaf_vector":"0"}},{"base_weights":[4.47928E-4,1.7265332E-1,-6.6526175E-2,1.0104353E-1,-7.7318847E-3,-1.428566E-1,1.177879E-1,-2.3108141E-1,6.693998E-2,-2.5350945E-2,7.2975345E-2,2.4930747E-2,-3.1360537E-1,6.517044E-2,-3.494925E-2,-1.1647142E-1,-9.730939E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,false,true,true,false,true,false,false,false,false,false,false,false,false],"id":50,"left_children":[1,3,5,-1,-1,7,9,11,13,-1,-1,-1,15,-1,-1,-1,-1],"loss_changes":[1.7304038E-1,1.4550528E-1,1.688699E-1,0E0,0E0,1.6528535E-1,1.1479147E-1,1.7962107E-1,1.1274567E-1,0E0,0E0,0E0,9.023163E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,5,5,6,6,7,7,8,8,12,12],"right_children":[2,4,6,-1,-1,8,10,12,14,-1,-1,-1,16,-1,-1,-1,-1],"split_conditions":[-6.383619E-1,-2.0705372E-1,3.1922126E0,1.0104353E-1,-7.7318847E-3,3.932939E0,-1.0445819E-1,-4.9166307E0,8.540317E-2,-2.5350945E-2,7.2975345E-2,2.4930747E-2,4.774607E-1,6.517044E-2,-3.494925E-2,-1.1647142E-1,-9.730939E-3],"split_indices":[403,219,6,0,0,16,298,14,260,0,0,0,184,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.3007647E1,3.2005095E0,9.8071375E0,1.3698307E0,1.8306788E0,7.094149E0,2.7129884E0,4.9226108E0,2.1715384E0,1.1634948E0,1.5494936E0,1.022553E0,3.9000578E0,1.0923252E0,1.0792131E0,2.781061E0,1.1189969E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"17","size_leaf_vector":"0"}},{"base_weights":[5.20288E-3,-1.0251859E-1,1.0699434E-1,1.06498934E-1,-2.3769873E-1,2.6905906E-1,-4.1704513E-2,-2.669686E-2,7.509074E-2,1.1065329E-1,6.970384E-3,-9.661866E-2,4.264782E-4,-1.9404507E-1,5.8273274E-2,-1.4218548E-4,-8.226967E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,false,true,false,false,false,false,false,false,false,true,false,false,false],"id":51,"left_children":[1,3,5,7,11,9,13,-1,-1,-1,-1,-1,-1,15,-1,-1,-1],"loss_changes":[1.6132124E-1,2.279756E-1,1.9290935E-1,1.20878E-1,9.456545E-2,8.321306E-2,2.1013911E-1,0E0,0E0,0E0,0E0,0E0,0E0,5.288616E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5,4,4,6,6,13,13],"right_children":[2,4,6,8,12,10,14,-1,-1,-1,-1,-1,-1,16,-1,-1,-1],"split_conditions":[-1.0163781E-1,-1.4197664E0,-2.431308E0,7.595253E-2,2.941962E0,2.062687E-1,2.042634E0,-2.669686E-2,7.509074E-2,1.1065329E-1,6.970384E-3,-9.661866E-2,4.264782E-4,-1.7381658E-1,5.8273274E-2,-1.4218548E-4,-8.226967E-2],"split_indices":[400,25,61,113,4,269,16,0,0,0,0,0,0,219,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.2711864E1,6.172564E0,6.5393004E0,2.5075397E0,3.6650243E0,2.7417188E0,3.7975817E0,1.224335E0,1.2832048E0,1.5941733E0,1.1475455E0,2.4527965E0,1.2122277E0,2.4155657E0,1.3820158E0,1.0021925E0,1.4133732E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"17","size_leaf_vector":"0"}},{"base_weights":[7.1759825E-3,5.966847E-2,-6.0796265E-2,-7.3500074E-2,1.6296792E-1,9.349159E-2,-2.2416005E-1,-2.7945582E-2,2.522244E-1,6.429947E-2,-1.4711105E-2,-2.5389437E-3,-9.624846E-2,3.4432125E-1,-3.8335475E-3,1.2675908E-1,2.2369713E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,true,true,false,false,false,false,false,false,false,true,false,false,false],"id":52,"left_children":[1,3,-1,5,7,9,11,-1,13,-1,-1,-1,-1,15,-1,-1,-1],"loss_changes":[1.5985768E-1,1.7022362E-1,0E0,1.6472597E-1,1.7869692E-1,6.9344275E-2,6.6060424E-2,0E0,1.3385639E-1,0E0,0E0,0E0,0E0,5.2294225E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,3,3,4,4,5,5,6,6,8,8,13,13],"right_children":[2,4,-1,6,8,10,12,-1,14,-1,-1,-1,-1,16,-1,-1,-1],"split_conditions":[7.6444206E0,-9.9897707E-1,-6.0796265E-2,-1.5112054E-1,-5.0228133E0,-2.2672982E0,-1.3470134E-1,-2.7945582E-2,3.9249587E0,6.429947E-2,-1.4711105E-2,-2.5389437E-3,-9.624846E-2,2.6334193E-1,-3.8335475E-3,1.2675908E-1,2.2369713E-2],"split_indices":[75,161,0,87,5,97,410,0,117,0,0,0,0,187,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.2415543E1,1.0503542E1,1.9120004E0,4.7144203E0,5.7891216E0,2.415989E0,2.2984314E0,1.4848182E0,4.3043036E0,1.0348411E0,1.3811479E0,1.0478649E0,1.2505665E0,2.972095E0,1.3322085E0,1.8650302E0,1.1070648E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"17","size_leaf_vector":"0"}},{"base_weights":[6.0952823E-3,7.0671804E-2,-1.6608748E-1,-4.6833266E-2,2.1433374E-1,4.009534E-3,-9.445592E-2,-1.506169E-1,5.280349E-2,-2.3845406E-2,9.77003E-2,5.875643E-2,-9.417596E-2,5.862912E-2,-3.188922E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,true,false,false,false,true,false,false,false,true,false,false,false],"id":53,"left_children":[1,3,5,7,9,-1,-1,11,-1,-1,-1,13,-1,-1,-1],"loss_changes":[1.5801868E-1,1.8339197E-1,1.0797195E-1,1.7613497E-1,1.8289192E-1,0E0,0E0,1.8778571E-1,0E0,0E0,0E0,8.87333E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,7,7,11,11],"right_children":[2,4,6,8,10,-1,-1,12,-1,-1,-1,14,-1,-1,-1],"split_conditions":[6.333127E0,5.617125E-1,1.7578545E0,3.7927504E0,-4.4699183E0,4.009534E-3,-9.445592E-2,-4.3729258E-1,5.280349E-2,-2.3845406E-2,9.77003E-2,-7.7842556E-2,-9.417596E-2,5.862912E-2,-3.188922E-2],"split_indices":[110,158,61,73,48,0,0,98,0,0,0,301,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.21559925E1,9.269176E0,2.8868175E0,5.4695134E0,3.799662E0,1.7209882E0,1.1658293E0,3.9527507E0,1.5167627E0,1.1227393E0,2.6769228E0,2.0125842E0,1.9401665E0,1.0002663E0,1.0123179E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"15","size_leaf_vector":"0"}},{"base_weights":[1.9050168E-3,-6.126239E-2,1.6203567E-1,6.5114886E-2,-1.8616134E-1,-1.1044815E-2,7.876789E-2,6.8934776E-2,-9.111155E-2,-2.987169E-1,1.5721442E-2,2.768046E-2,-6.292871E-2,-1.0839103E-1,-2.3374062E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,true,false,false,false,true,true,false,false,false,false,false],"id":54,"left_children":[1,3,5,7,9,-1,-1,-1,11,13,-1,-1,-1,-1,-1],"loss_changes":[1.4015968E-1,1.6931313E-1,8.795541E-2,1.6733666E-1,1.5710372E-1,0E0,0E0,0E0,9.8475054E-2,2.3373187E-2,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,8,8,9,9],"right_children":[2,4,6,8,10,-1,-1,-1,12,14,-1,-1,-1,-1,-1],"split_conditions":[4.8296633E0,8.5754335E-2,-1.4768566E0,-6.921234E-1,3.5339298E0,-1.1044815E-2,7.876789E-2,6.8934776E-2,-3.7720084E-1,2.6950642E-1,1.5721442E-2,2.768046E-2,-6.292871E-2,-1.0839103E-1,-2.3374062E-2],"split_indices":[5,147,38,95,36,0,0,0,80,296,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.1875187E1,8.9586525E0,2.916534E0,4.690895E0,4.2677574E0,1.1921122E0,1.7244217E0,2.0545273E0,2.6363678E0,2.7283752E0,1.5393822E0,1.1230311E0,1.5133367E0,1.6300312E0,1.098344E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"15","size_leaf_vector":"0"}},{"base_weights":[6.9044854E-6,-8.903071E-2,1.1246915E-1,-2.231805E-1,7.429666E-2,-4.501092E-2,8.056866E-2,-3.0727616E-1,3.1231663E-3,7.082923E-2,-4.6615597E-2,-6.1159097E-2,3.8438153E-2,-2.7195293E-2,-1.1086016E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,true,true,false,false,true,false,false,false,false,false,false,false],"id":55,"left_children":[1,3,5,7,9,11,-1,13,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[1.3673843E-1,1.8357287E-1,1.6041385E-1,9.0306655E-2,1.9229594E-1,1.3228728E-1,0E0,1.0916382E-2,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,5,5,7,7],"right_children":[2,4,6,8,10,12,-1,14,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[1.5378008E0,8.1120896E-1,1.3970428E0,5.4800105E-1,7.670407E-1,1.3853186E0,8.056866E-2,-1.2786137E-1,3.1231663E-3,7.082923E-2,-4.6615597E-2,-6.1159097E-2,3.8438153E-2,-2.7195293E-2,-1.1086016E-1],"split_indices":[78,36,32,348,37,73,0,428,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.1655601E1,6.6215754E0,5.034025E0,3.4343185E0,3.187257E0,2.860125E0,2.1739001E0,2.2932687E0,1.1410497E0,1.8535684E0,1.3336886E0,1.3990477E0,1.4610773E0,1.0602384E0,1.2330303E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"15","size_leaf_vector":"0"}},{"base_weights":[-4.3868725E-3,-8.2291655E-2,1.1818929E-1,-1.6804439E-1,9.429748E-2,-3.861165E-2,7.381781E-2,-2.644251E-1,1.8076714E-2,6.4921975E-2,-2.1768348E-2,3.2877307E-2,-4.8923407E-2,-5.584962E-3,-1.0151302E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,true,true,true,false,false,false,false,false,false,false,false,false],"id":56,"left_children":[1,3,5,7,9,11,-1,13,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[1.2831242E-1,1.3975008E-1,1.1448577E-1,1.4503193E-1,8.3248176E-2,7.720153E-2,0E0,7.206705E-2,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,5,5,7,7],"right_children":[2,4,6,8,10,12,-1,14,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[3.959839E-1,3.9791207E0,1.7952664E0,2.853623E0,6.7946744E-1,-1.6151662E0,7.381781E-2,-3.9497665E-1,1.8076714E-2,6.4921975E-2,-2.1768348E-2,3.2877307E-2,-4.8923407E-2,-5.584962E-3,-1.0151302E-1],"split_indices":[177,97,133,52,121,85,0,479,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.1418378E1,7.1822557E0,4.2361217E0,4.8671346E0,2.3151214E0,2.2163537E0,2.019768E0,3.3110743E0,1.5560601E0,1.1653553E0,1.1497661E0,1.0662627E0,1.1500909E0,1.0552517E0,2.2558227E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"15","size_leaf_vector":"0"}},{"base_weights":[-2.3605146E-3,1.1807933E-1,-8.7140106E-2,-2.695317E-2,2.6043716E-1,-1.557014E-1,4.1570727E-2,1.1030431E-1,1.5838481E-2,-2.5901273E-1,1.4881429E-2,-1.0615735E-1,2.9590956E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,false,true,true,false,false,false,true,false,false,false],"id":57,"left_children":[1,3,5,-1,7,9,-1,-1,-1,11,-1,-1,-1],"loss_changes":[1.3507582E-1,1.8540585E-1,1.4146984E-1,0E0,5.8647558E-2,1.4888036E-1,0E0,0E0,0E0,1.16392195E-1,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,4,4,5,5,9,9],"right_children":[2,4,6,-1,8,10,-1,-1,-1,12,-1,-1,-1],"split_conditions":[-1.2793412E0,-1.1932282E0,6.9316564E0,-2.695317E-2,7.3667336E-1,2.1515646E0,4.1570727E-2,1.1030431E-1,1.5838481E-2,3.4923568E0,1.4881429E-2,-1.0615735E-1,2.9590956E-3],"split_indices":[86,101,161,0,64,47,0,0,0,41,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.1220972E1,4.4733076E0,6.7476645E0,1.9679177E0,2.50539E0,5.413447E0,1.3342174E0,1.143861E0,1.3615289E0,3.427244E0,1.9862028E0,2.2999012E0,1.1273427E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"13","size_leaf_vector":"0"}},{"base_weights":[2.5747654E-3,-1.4820728E-1,6.513848E-2,-7.412973E-2,1.5295964E-2,1.3235638E-1,-3.9571356E-2,1.9741324E-1,-1.9662153E-2,-6.73015E-3,2.7192703E-1,1.058056E-1,2.0988347E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,false,false,true,false,true,false,false,false,false,false],"id":58,"left_children":[1,3,5,-1,-1,7,-1,9,-1,-1,11,-1,-1],"loss_changes":[1.229953E-1,8.80181E-2,1.3912699E-1,0E0,0E0,1.07549414E-1,0E0,1.00782976E-1,0E0,0E0,4.697624E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,5,5,7,7,10,10],"right_children":[2,4,6,-1,-1,8,-1,10,-1,-1,12,-1,-1],"split_conditions":[-5.332983E0,1.9460897E0,6.333127E0,-7.412973E-2,1.5295964E-2,7.683147E-1,-3.9571356E-2,-5.6079876E-1,-1.9662153E-2,-6.73015E-3,1.324436E-2,1.058056E-1,2.0988347E-2],"split_indices":[19,5,110,0,0,238,0,376,0,0,292,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.1014787E1,2.828659E0,8.186128E0,1.7295237E0,1.0991354E0,6.348672E0,1.8374557E0,4.779799E0,1.5688727E0,1.3868724E0,3.3929267E0,1.8906624E0,1.5022643E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"13","size_leaf_vector":"0"}},{"base_weights":[2.259844E-3,9.766019E-2,-9.652829E-2,-3.427112E-2,1.8126859E-1,-2.3698133E-1,3.3303346E-2,2.778159E-1,-1.3290242E-2,-8.6668506E-2,-2.0301353E-2,6.412842E-2,-3.2459013E-2,1.07142076E-1,1.0885807E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,true,true,true,true,false,false,false,false,false,false,false],"id":59,"left_children":[1,3,5,-1,7,9,11,13,-1,-1,-1,-1,-1,-1,-1],"loss_changes":[1.21180296E-1,1.3638572E-1,1.2329224E-1,0E0,1.22692645E-1,8.251995E-3,1.294892E-1,5.7679802E-2,0E0,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,4,4,5,5,6,6,7,7],"right_children":[2,4,6,-1,8,10,12,14,-1,-1,-1,-1,-1,-1,-1],"split_conditions":[3.3998775E-1,-5.6151323E0,-1.9369357E0,-3.427112E-2,4.538559E-1,1.5412903E0,-3.6964476E-1,4.0927086E0,-1.3290242E-2,-8.6668506E-2,-2.0301353E-2,6.412842E-2,-3.2459013E-2,1.07142076E-1,1.0885807E-2],"split_indices":[159,48,10,0,339,137,348,25,0,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.0857843E1,5.5294423E0,5.328401E0,1.4608148E0,4.0686274E0,2.1630726E0,3.1653285E0,2.6869495E0,1.3816779E0,1.114913E0,1.0481596E0,1.1667134E0,1.9986151E0,1.662338E0,1.0246115E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"15","size_leaf_vector":"0"}},{"base_weights":[1.4812831E-3,8.68196E-2,-1.0604568E-1,-2.9083263E-2,1.6006042E-1,-2.0039488E-1,3.0375866E-2,2.4541627E-1,-2.0814147E-2,-8.435884E-2,4.8805214E-3,1.01977214E-1,5.9850677E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,false,true,true,false,true,false,false,false,false,false],"id":60,"left_children":[1,3,5,-1,7,9,-1,11,-1,-1,-1,-1,-1],"loss_changes":[1.1632264E-1,1.1051009E-1,1.3011709E-1,0E0,1.2324746E-1,7.800828E-2,0E0,8.214983E-2,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,4,4,5,5,7,7],"right_children":[2,4,6,-1,8,10,-1,12,-1,-1,-1,-1,-1],"split_conditions":[2.1315716E-1,-4.074327E0,5.35488E-1,-2.9083263E-2,3.6965723E0,4.3108773E-1,3.0375866E-2,1.1287964E-1,-2.0814147E-2,-8.435884E-2,4.8805214E-3,1.01977214E-1,5.9850677E-3],"split_indices":[280,90,271,0,85,328,0,210,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.0673256E1,6.0579605E0,4.6152954E0,1.6341594E0,4.423801E0,3.1946125E0,1.4206829E0,3.1735573E0,1.2502435E0,2.109907E0,1.0847056E0,1.878502E0,1.2950553E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"13","size_leaf_vector":"0"}},{"base_weights":[1.984769E-3,-9.656703E-2,9.378719E-2,3.9272975E-2,-2.2518042E-1,-2.5640236E-2,1.8996188E-1,4.444716E-2,-3.9650977E-2,-5.330162E-3,-9.3591124E-2,8.798034E-2,-1.7277945E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,false,false,true,false,false,false,false,false,false],"id":61,"left_children":[1,3,5,7,9,-1,11,-1,-1,-1,-1,-1,-1],"loss_changes":[1.13156095E-1,1.1439844E-1,1.2767947E-1,9.0328336E-2,5.2996695E-2,0E0,1.3241853E-1,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,6,6],"right_children":[2,4,6,8,10,-1,12,-1,-1,-1,-1,-1,-1],"split_conditions":[-1.13829784E-1,2.0587774E-1,-3.4886904E0,3.5996762E-1,-1.02570586E-1,-2.5640236E-2,3.271019E0,4.444716E-2,-3.9650977E-2,-5.330162E-3,-9.3591124E-2,8.798034E-2,-1.7277945E-2],"split_indices":[400,212,59,260,308,0,61,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.0506131E1,5.041778E0,5.464353E0,2.789832E0,2.251946E0,1.9469229E0,3.51743E0,1.7892761E0,1.000556E0,1.0197154E0,1.2322305E0,2.3514864E0,1.1659436E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"13","size_leaf_vector":"0"}},{"base_weights":[-2.8557675E-3,-5.772262E-2,1.6470811E-1,-1.4253485E-1,9.1089286E-2,7.8028485E-2,-3.4949481E-3,-7.520909E-3,-8.6839765E-2,-2.7043073E-2,7.0107594E-2,1.4427514E-1,-6.1684582E-2,6.475998E-2,2.7774295E-4],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,false,true,false,false,false,false,false,false,false,false,false,false],"id":62,"left_children":[1,3,5,7,9,-1,-1,11,-1,-1,-1,13,-1,-1,-1],"loss_changes":[1.1377452E-1,1.3015904E-1,5.4455988E-2,1.2287253E-1,1.2393919E-1,0E0,0E0,1.593179E-1,0E0,0E0,0E0,3.0883223E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,7,7,11,11],"right_children":[2,4,6,8,10,-1,-1,12,-1,-1,-1,14,-1,-1,-1],"split_conditions":[6.313111E0,3.2208195E0,-1.2516917E-1,-2.097207E0,-3.511431E-2,7.8028485E-2,-3.4949481E-3,2.2912774E0,-8.6839765E-2,-2.7043073E-2,7.0107594E-2,1.4130208E0,-6.1684582E-2,6.475998E-2,2.7774295E-4],"split_indices":[154,88,464,88,387,0,0,81,0,0,0,120,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.034115E1,8.284127E0,2.0570228E0,5.303621E0,2.9805064E0,1.0268285E0,1.0301943E0,3.3117044E0,1.9919164E0,1.4744747E0,1.5060318E0,2.0287757E0,1.2829287E0,1.0156584E0,1.0131173E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"15","size_leaf_vector":"0"}},{"base_weights":[1.0053536E-3,8.537702E-2,-1.1349313E-1,1.8912661E-1,-7.420075E-2,-2.0829202E-1,3.5403505E-2,-6.3111656E-3,2.6846045E-1,-6.241843E-2,2.0827632E-2,-8.544438E-2,-5.4715322E-3,9.975168E-2,2.9649109E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,true,true,false,false,false,true,false,false,false,false,false,false],"id":63,"left_children":[1,3,5,7,9,11,-1,-1,13,-1,-1,-1,-1,-1,-1],"loss_changes":[1.1737989E-1,1.2974146E-1,1.377962E-1,8.096096E-2,8.119294E-2,5.3293854E-2,0E0,0E0,3.5530329E-3,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,5,5,8,8],"right_children":[2,4,6,8,10,12,-1,-1,14,-1,-1,-1,-1,-1,-1],"split_conditions":[1.3989339E0,2.7233639E0,5.495457E0,-4.0114326E0,-1.8500423E0,2.1471457E-1,3.5403505E-2,-6.3111656E-3,-2.4530811E0,-6.241843E-2,2.0827632E-2,-8.544438E-2,-5.4715322E-3,9.975168E-2,2.9649109E-2],"split_indices":[117,125,16,2,105,188,0,0,61,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[1.0147557E1,5.9888396E0,4.1587176E0,3.517055E0,2.4717846E0,3.021651E0,1.1370667E0,1.1651839E0,2.351871E0,1.0471766E0,1.424608E0,1.7987903E0,1.2228607E0,1.0102593E0,1.3416117E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"15","size_leaf_vector":"0"}},{"base_weights":[3.0733668E-3,6.517164E-2,-1.4188163E-1,1.2760216E-1,-5.1590875E-2,-7.9886906E-2,1.2085465E-2,2.3508042E-1,-4.184269E-2,3.0380234E-1,-6.217925E-4,-6.2289897E-2,3.8338803E-2,3.5040844E-2,1.1476595E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,false,false,false,true,true,false,false,false,false,false,false],"id":64,"left_children":[1,3,5,7,-1,-1,-1,9,11,13,-1,-1,-1,-1,-1],"loss_changes":[1.0814922E-1,1.4596927E-1,9.2541665E-2,1.4106178E-1,0E0,0E0,0E0,7.494804E-2,1.2906235E-1,7.48232E-3,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,7,7,8,8,9,9],"right_children":[2,4,6,8,-1,-1,-1,10,12,14,-1,-1,-1,-1,-1],"split_conditions":[5.986258E-1,7.726712E0,-1.8168974E-1,1.680795E0,-5.1590875E-2,-7.9886906E-2,1.2085465E-2,5.2126515E-1,-1.5717621E0,7.515925E-1,-6.217925E-4,-6.2289897E-2,3.8338803E-2,3.5040844E-2,1.1476595E-1],"split_indices":[425,69,11,134,0,0,0,314,92,156,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[9.9874E0,7.3773623E0,2.6100378E0,6.2055774E0,1.1717846E0,1.2764865E0,1.3335513E0,3.5600767E0,2.6455007E0,2.5423257E0,1.017751E0,1.2246556E0,1.420845E0,1.4892317E0,1.053094E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"15","size_leaf_vector":"0"}},{"base_weights":[2.920282E-3,1.4806965E-1,-5.4860156E-2,-7.104197E-3,8.220664E-2,-1.8856877E-1,5.750787E-2,4.7921236E-3,-8.479503E-2,-5.0312866E-2,2.0782441E-1,8.896445E-2,-2.3417312E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,false,false,false,true,false,false,false,false,false,false],"id":65,"left_children":[1,3,5,-1,-1,7,9,-1,-1,-1,11,-1,-1],"loss_changes":[9.871341E-2,7.886825E-2,1.3788691E-1,0E0,0E0,8.33153E-2,2.1633764E-1,0E0,0E0,0E0,7.236658E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,5,5,6,6,10,10],"right_children":[2,4,6,-1,-1,8,10,-1,-1,-1,12,-1,-1],"split_conditions":[-5.6589746E0,1.7211767E-1,-2.0525303E0,-7.104197E-3,8.220664E-2,-3.1324092E-1,-2.1855652E0,4.7921236E-3,-8.479503E-2,-5.0312866E-2,2.1984572E0,8.896445E-2,-2.3417312E-3],"split_indices":[57,400,94,0,0,265,3,0,0,0,64,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[9.801538E0,2.345875E0,7.4556637E0,1.3360323E0,1.0098428E0,3.094879E0,4.3607845E0,1.2365985E0,1.8582804E0,1.6991941E0,2.6615906E0,1.6198289E0,1.0417616E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"13","size_leaf_vector":"0"}},{"base_weights":[5.1628826E-3,-5.9192184E-2,1.3900165E-1,-1.6744861E-1,8.327855E-2,-1.3613095E-2,7.235982E-2,-2.6182756E-1,1.9236125E-2,-2.7980577E-2,6.589588E-2,-1.0798655E-1,-6.4802947E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,true,true,false,false,true,false,false,false,false,false],"id":66,"left_children":[1,3,5,7,9,-1,-1,11,-1,-1,-1,-1,-1],"loss_changes":[9.9700876E-2,1.3537328E-1,8.155153E-2,1.21821076E-1,1.1841021E-1,0E0,0E0,7.7346474E-2,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4,7,7],"right_children":[2,4,6,8,10,-1,-1,12,-1,-1,-1,-1,-1],"split_conditions":[4.8019023E0,1.551867E0,-2.1069776E-1,3.5659223E0,-4.4177806E-1,-1.3613095E-2,7.235982E-2,1.0310843E0,1.9236125E-2,-2.7980577E-2,6.589588E-2,-1.0798655E-1,-6.4802947E-3],"split_indices":[70,3,264,113,78,0,0,152,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[9.620097E0,6.8730116E0,2.7470853E0,3.805831E0,3.0671809E0,1.1779292E0,1.5691562E0,2.611012E0,1.1948189E0,1.4744611E0,1.5927198E0,1.499925E0,1.111087E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"13","size_leaf_vector":"0"}},{"base_weights":[3.5939326E-3,1.2733687E-1,-6.333304E-2,-1.904454E-2,6.821248E-2,-1.889422E-1,4.61947E-2,6.644299E-4,-9.485302E-2,-4.49943E-2,1.5454845E-1,-1.1265084E-2,7.698245E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,false,false,true,false,false,false,false,false,false],"id":67,"left_children":[1,3,5,-1,-1,7,9,-1,-1,-1,11,-1,-1],"loss_changes":[9.462943E-2,9.081137E-2,1.112422E-1,0E0,0E0,8.989674E-2,1.2487087E-1,0E0,0E0,0E0,7.9712786E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,5,5,6,6,10,10],"right_children":[2,4,6,-1,-1,8,10,-1,-1,-1,12,-1,-1],"split_conditions":[-3.5090895E0,-2.84872E-1,-7.8551006E-1,-1.904454E-2,6.821248E-2,1.0165196E-1,-3.415132E0,6.644299E-4,-9.485302E-2,-4.49943E-2,-9.372046E-2,-1.1265084E-2,7.698245E-2],"split_indices":[164,355,11,0,0,238,154,0,0,0,346,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[9.452488E0,3.0010846E0,6.451403E0,1.1578857E0,1.8431989E0,2.6673532E0,3.78405E0,1.458579E0,1.2087742E0,1.2096983E0,2.5743515E0,1.1124854E0,1.4618661E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"13","size_leaf_vector":"0"}},{"base_weights":[1.7860947E-4,1.2695158E-1,-6.4423315E-2,-2.7728159E-2,7.1830355E-2,5.162381E-2,-2.0787147E-1,6.5713555E-2,-5.8716673E-2,-8.1629984E-2,-7.6791523E-3,2.2953186E-2,-5.098552E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false,true,true,false,true,false,false,false,false],"id":68,"left_children":[1,3,5,-1,-1,7,9,-1,11,-1,-1,-1,-1],"loss_changes":[9.2499405E-2,1.1620098E-1,1.3530998E-1,0E0,0E0,1.0408327E-1,3.454481E-2,0E0,6.963634E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,5,5,6,6,8,8],"right_children":[2,4,6,-1,-1,8,10,-1,12,-1,-1,-1,-1],"split_conditions":[-3.6159916E0,-5.0562966E-1,2.246585E-1,-2.7728159E-2,7.1830355E-2,-5.0805867E-1,2.1711068E0,6.5713555E-2,1.2873737E0,-8.1629984E-2,-7.6791523E-3,2.2953186E-2,-5.098552E-2],"split_indices":[159,480,226,0,0,287,161,0,9,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[9.295856E0,2.812179E0,6.4836774E0,1.0136086E0,1.7985705E0,3.9380136E0,2.545664E0,1.173011E0,2.7650025E0,1.5179685E0,1.0276955E0,1.388811E0,1.3761915E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"13","size_leaf_vector":"0"}},{"base_weights":[-6.228297E-4,1.3897705E-1,-5.8249407E-2,7.101611E-2,-1.1098582E-2,-1.3437392E-1,1.0585657E-1,2.574918E-2,-7.6226115E-2,6.629388E-2,-1.8691018E-2,5.629519E-2,-4.0559467E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,false,true,true,false,false,false,false,false,false],"id":69,"left_children":[1,3,5,-1,-1,7,9,11,-1,-1,-1,-1,-1],"loss_changes":[8.974895E-2,6.484762E-2,1.12777665E-1,0E0,0E0,1.1829253E-1,7.29917E-2,1.1401662E-1,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,5,5,6,6,7,7],"right_children":[2,4,6,-1,-1,8,10,12,-1,-1,-1,-1,-1],"split_conditions":[-6.383619E-1,7.347642E-2,3.11203E0,7.101611E-2,-1.1098582E-2,8.099456E-1,1.0035095E0,-7.222166E-1,-7.6226115E-2,6.629388E-2,-1.8691018E-2,5.629519E-2,-4.0559467E-2],"split_indices":[403,249,71,0,0,83,106,147,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[9.150035E0,2.2610288E0,6.889006E0,1.2316815E0,1.0293473E0,4.829775E0,2.0592313E0,2.4019456E0,2.4278293E0,1.0359274E0,1.0233039E0,1.1147126E0,1.287233E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"13","size_leaf_vector":"0"}},{"base_weights":[4.2755776E-3,7.5324945E-2,-1.1099755E-1,-6.967051E-2,1.8737483E-1,-6.2896706E-2,1.9902762E-2,-5.273002E-2,3.2010704E-2,8.5791394E-2,-4.317743E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,true,true,false,false,false,false,false,false],"id":70,"left_children":[1,3,5,7,9,-1,-1,-1,-1,-1,-1],"loss_changes":[9.025014E-2,1.232334E-1,8.768955E-2,8.7713234E-2,8.5608706E-2,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4],"right_children":[2,4,6,8,10,-1,-1,-1,-1,-1,-1],"split_conditions":[2.2912774E0,-1.005744E0,2.3420334E0,3.533063E0,1.7179898E0,-6.2896706E-2,1.9902762E-2,-5.273002E-2,3.2010704E-2,8.5791394E-2,-4.317743E-3],"split_indices":[81,161,45,153,132,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[8.998587E0,5.7816067E0,3.21698E0,2.6851594E0,3.0964472E0,1.9499544E0,1.2670256E0,1.6787516E0,1.0064079E0,1.7996769E0,1.2967703E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"11","size_leaf_vector":"0"}},{"base_weights":[4.940964E-3,1.431274E-1,-5.27793E-2,2.1721441E-3,6.52663E-2,-1.2222866E-1,3.495103E-2,1.6173609E-2,-2.0701733E-1,-7.656433E-3,-8.780085E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,false,false,true,false,false,true,false,false],"id":71,"left_children":[1,3,5,-1,-1,7,-1,-1,9,-1,-1],"loss_changes":[8.630236E-2,3.0568928E-2,1.0470291E-1,0E0,0E0,9.993608E-2,0E0,0E0,5.8121383E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,5,5,8,8],"right_children":[2,4,6,-1,-1,8,-1,-1,10,-1,-1],"split_conditions":[-6.518854E-1,1.2529698E-1,6.1074585E-1,2.1721441E-3,6.52663E-2,-3.055019E-1,3.495103E-2,1.6173609E-2,-1.8508029E-1,-7.656433E-3,-8.780085E-2],"split_indices":[212,487,447,0,0,460,0,0,95,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[8.87295E0,2.1782913E0,6.6946583E0,1.1591771E0,1.0191143E0,4.9442205E0,1.7504377E0,1.7249519E0,3.2192688E0,1.4483017E0,1.7709671E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"11","size_leaf_vector":"0"}},{"base_weights":[5.675781E-3,5.1853333E-2,-4.8827283E-2,1.1648821E-1,-3.328377E-2,-2.8996343E-2,1.7553887E-1,1.958844E-3,2.6613006E-1,1.9936947E-2,1.0514329E-1],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,true,false,false,true,false,false,false,false],"id":72,"left_children":[1,3,-1,5,-1,-1,7,-1,9,-1,-1],"loss_changes":[8.445435E-2,1.0161538E-1,0E0,9.9043585E-2,0E0,0E0,8.1882685E-2,0E0,3.616929E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,3,3,6,6,8,8],"right_children":[2,4,-1,6,-1,-1,8,-1,10,-1,-1],"split_conditions":[7.6444206E0,6.742075E-1,-4.8827283E-2,-6.839275E0,-3.328377E-2,-2.8996343E-2,-1.661406E0,1.958844E-3,4.1179466E-1,1.9936947E-2,1.0514329E-1],"split_indices":[75,280,0,56,0,0,76,0,88,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[8.773254E0,7.4287405E0,1.3445138E0,5.5211782E0,1.9075623E0,1.0596344E0,4.461544E0,1.9310325E0,2.5305114E0,1.2824627E0,1.2480487E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"11","size_leaf_vector":"0"}},{"base_weights":[6.0107615E-3,-9.560634E-2,8.131035E-2,-6.726098E-2,2.6704237E-2,1.7737469E-1,-2.5565278E-2,8.422682E-2,-3.5201162E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false,false,false,false,false],"id":73,"left_children":[1,3,5,-1,-1,7,-1,-1,-1],"loss_changes":[8.173429E-2,1.2854184E-1,1.125758E-1,0E0,0E0,8.517188E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,5,5],"right_children":[2,4,6,-1,-1,8,-1,-1,-1],"split_conditions":[-1.3897004E0,1.3853186E0,2.0011115E0,-6.726098E-2,2.6704237E-2,1.2873737E0,-2.5565278E-2,8.422682E-2,-3.5201162E-3],"split_indices":[36,73,117,0,0,9,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[8.665859E0,3.5735967E0,5.0922627E0,1.9800189E0,1.5935779E0,3.1880498E0,1.9042131E0,1.7478882E0,1.4401616E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[5.8964975E-3,-1.13884956E-1,7.103374E-2,-6.5198265E-2,2.3570428E-2,-3.2671314E-2,1.3296105E-1,-1.1013826E-2,2.0028333E-1,5.072389E-3,8.472428E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false,false,false,false,true,false,false],"id":74,"left_children":[1,3,5,-1,-1,-1,7,-1,9,-1,-1],"loss_changes":[8.2663886E-2,9.174819E-2,9.0332314E-2,0E0,0E0,0E0,7.0344016E-2,0E0,5.701822E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,6,6,8,8],"right_children":[2,4,6,-1,-1,-1,8,-1,10,-1,-1],"split_conditions":[-3.208375E0,1.5817978E0,-6.001896E0,-6.5198265E-2,2.3570428E-2,-3.2671314E-2,-4.282502E0,-1.1013826E-2,-1.0050347E0,5.072389E-3,8.472428E-2],"split_indices":[153,103,135,0,0,0,4,0,19,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[8.558061E0,2.7509427E0,5.807118E0,1.7051713E0,1.0457714E0,1.2926284E0,4.5144897E0,1.4115646E0,3.102925E0,1.3328686E0,1.7700565E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"11","size_leaf_vector":"0"}},{"base_weights":[4.7862264E-3,-8.610732E-2,9.482117E-2,-6.391791E-2,4.4709023E-2,-2.5665095E-2,1.8705101E-1,-3.6125727E-2,5.3551294E-2,-3.7514165E-3,8.2985416E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false,false,false,false,false,false,false],"id":75,"left_children":[1,3,5,-1,7,-1,9,-1,-1,-1,-1],"loss_changes":[8.553917E-2,9.636693E-2,1.02936104E-1,0E0,9.6654475E-2,0E0,7.087019E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,4,4,6,6],"right_children":[2,4,6,-1,8,-1,10,-1,-1,-1,-1],"split_conditions":[-8.406782E-2,-1.8037028E0,-3.2823224E0,-6.391791E-2,-8.3916736E-1,-2.5665095E-2,-4.148978E0,-3.6125727E-2,5.3551294E-2,-3.7514165E-3,8.2985416E-2],"split_indices":[16,94,3,0,38,0,55,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[8.454813E0,4.229053E0,4.22576E0,1.8271646E0,2.4018884E0,1.4542142E0,2.771546E0,1.1198103E0,1.282078E0,1.1251326E0,1.6464133E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"11","size_leaf_vector":"0"}},{"base_weights":[4.3105353E-3,7.8477636E-2,-1.07922696E-1,1.9588411E-1,-4.1363265E-2,-7.2796725E-2,1.0993439E-2,8.627887E-2,1.5075092E-3,-4.5395166E-2,3.5283644E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,true,true,false,false,false,false,false,false],"id":76,"left_children":[1,3,5,7,9,-1,-1,-1,-1,-1,-1],"loss_changes":[8.617345E-2,9.533158E-2,8.94395E-2,5.6422666E-2,8.5857674E-2,0E0,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,4,4],"right_children":[2,4,6,8,10,-1,-1,-1,-1,-1,-1],"split_conditions":[3.0724077E0,-1.6616076E-1,-1.6451392E0,3.239441E-1,1.5698647E0,-7.2796725E-2,1.0993439E-2,8.627887E-2,1.5075092E-3,-4.5395166E-2,3.5283644E-2],"split_indices":[25,425,94,83,154,0,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[8.334912E0,5.1996136E0,3.1352987E0,2.305961E0,2.8936527E0,1.2716553E0,1.8636434E0,1.2151855E0,1.0907754E0,1.7390361E0,1.1546166E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"11","size_leaf_vector":"0"}},{"base_weights":[4.838134E-3,-7.82913E-2,1.0308668E-1,2.7936084E-2,-1.8962736E-1,-2.5322845E-2,1.8350531E-1,-7.844846E-2,-4.643869E-3,7.208522E-2,9.675214E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,true,false,true,false,false,false,false],"id":77,"left_children":[1,3,5,-1,7,-1,9,-1,-1,-1,-1],"loss_changes":[8.340913E-2,1.23861805E-1,8.574428E-2,0E0,4.2297646E-2,0E0,2.2960164E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,4,4,6,6],"right_children":[2,4,6,-1,8,-1,10,-1,-1,-1,-1],"split_conditions":[1.5412903E0,-3.291986E0,-5.081229E-1,2.7936084E-2,1.9078732E-2,-2.5322845E-2,1.4477749E0,-7.844846E-2,-4.643869E-3,7.208522E-2,9.675214E-3],"split_indices":[137,63,360,0,108,0,134,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[8.224338E0,4.5649776E0,3.6593597E0,1.8619516E0,2.703026E0,1.0835156E0,2.575844E0,1.5583543E0,1.1446718E0,1.444859E0,1.130985E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"11","size_leaf_vector":"0"}},{"base_weights":[3.5122714E-3,-7.033634E-2,1.0188543E-1,-1.4550093E-1,4.0181622E-2,1.9253343E-1,-2.401791E-2,4.9821907E-3,-7.171102E-2,1.2367969E-2,7.8213796E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,false,false,false,false,false,false,false,false],"id":78,"left_children":[1,3,5,7,-1,9,-1,-1,-1,-1,-1],"loss_changes":[7.3491916E-2,1.0864356E-1,8.6717725E-2,7.559944E-2,0E0,2.2049427E-2,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,5,5],"right_children":[2,4,6,8,-1,10,-1,-1,-1,-1,-1],"split_conditions":[1.7952664E0,5.089202E0,1.0829055E0,-1.8835588E0,4.0181622E-2,4.2577863E-1,-2.401791E-2,4.9821907E-3,-7.171102E-2,1.2367969E-2,7.8213796E-2],"split_indices":[133,31,142,22,0,112,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[8.12983E0,4.806558E0,3.3232722E0,3.7240028E0,1.082555E0,2.1793098E0,1.1439624E0,1.6634694E0,2.0605335E0,1.175425E0,1.0038848E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"11","size_leaf_vector":"0"}},{"base_weights":[6.9080917E-3,7.326982E-2,-1.0098994E-1,-1.7407788E-2,6.389055E-2,-5.685274E-2,2.1381212E-2,-5.0290134E-2,1.06932E-1,6.078509E-2,-1.140732E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,true,false,false,false,false,true,false,false],"id":79,"left_children":[1,3,5,7,-1,-1,-1,-1,9,-1,-1],"loss_changes":[7.207375E-2,8.189353E-2,7.236766E-2,1.05104394E-1,0E0,0E0,0E0,0E0,5.1591694E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,8,8],"right_children":[2,4,6,8,-1,-1,-1,-1,10,-1,-1],"split_conditions":[2.2912774E0,4.578299E0,3.8656574E-1,-1.5896711E0,6.389055E-2,-5.685274E-2,2.1381212E-2,-5.0290134E-2,-7.560509E-1,6.078509E-2,-1.140732E-2],"split_indices":[81,16,413,56,0,0,0,0,37,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[8.032335E0,5.1721716E0,2.8601637E0,3.667202E0,1.5049698E0,1.823167E0,1.0369967E0,1.5030464E0,2.1641555E0,1.0640217E0,1.1001338E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"11","size_leaf_vector":"0"}},{"base_weights":[8.02261E-3,-5.2328747E-2,1.2932843E-1,-1.18292645E-1,3.652824E-2,-9.743574E-3,6.841647E-2,2.1189233E-2,-1.9473813E-1,-7.641627E-2,-1.7280574E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false,false,false,false,true,false,false],"id":80,"left_children":[1,3,5,7,-1,-1,-1,-1,9,-1,-1],"loss_changes":[7.214666E-2,9.106131E-2,5.940392E-2,9.09966E-2,0E0,0E0,0E0,0E0,4.4710636E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,8,8],"right_children":[2,4,6,8,-1,-1,-1,-1,10,-1,-1],"split_conditions":[5.0954676E-1,5.0261116E0,-1.4648776E-1,-5.7606524E-1,3.652824E-2,-9.743574E-3,6.841647E-2,2.1189233E-2,5.7021576E-1,-7.641627E-2,-1.7280574E-3],"split_indices":[487,5,264,496,0,0,0,0,221,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[7.9303827E0,5.6754084E0,2.2549744E0,4.348307E0,1.3271012E0,1.108777E0,1.1461973E0,1.2745395E0,3.0737677E0,2.069127E0,1.0046406E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"11","size_leaf_vector":"0"}},{"base_weights":[9.9134E-3,-7.315242E-2,9.996619E-2,-1.6335377E-1,2.868536E-2,-1.5978863E-2,5.8988053E-2,6.266263E-3,-7.945791E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,false,false,false,false,false],"id":81,"left_children":[1,3,5,7,-1,-1,-1,-1,-1],"loss_changes":[7.3404916E-2,9.437706E-2,7.952283E-2,7.549021E-2,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3],"right_children":[2,4,6,8,-1,-1,-1,-1,-1],"split_conditions":[1.0041963E0,3.7718365E0,-2.1855652E0,-8.6186066E-2,2.868536E-2,-1.5978863E-2,5.8988053E-2,6.266263E-3,-7.945791E-2],"split_indices":[78,36,3,485,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[7.8354836E0,4.173483E0,3.6620007E0,2.7407496E0,1.4327333E0,1.5901779E0,2.0718226E0,1.2557275E0,1.4850221E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[6.7367125E-3,-4.571481E-2,4.6253484E-2,9.36837E-2,-1.2301825E-1,5.635696E-2,-1.38993515E-2,1.8735897E-2,-1.876086E-1,-7.9434305E-2,-2.9041027E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,true,true,false,false,false,false,false,false],"id":82,"left_children":[1,3,-1,5,7,-1,-1,-1,9,-1,-1],"loss_changes":[7.4713364E-2,8.938138E-2,0E0,5.017283E-2,7.257297E-2,0E0,0E0,0E0,5.2602753E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,3,3,4,4,8,8],"right_children":[2,4,-1,6,8,-1,-1,-1,10,-1,-1],"split_conditions":[6.313111E0,-4.290621E-1,4.6253484E-2,-9.814325E-2,-6.3342524E-1,5.635696E-2,-1.38993515E-2,1.8735897E-2,3.1319487E0,-7.9434305E-2,-2.9041027E-3],"split_indices":[154,280,0,425,439,0,0,0,124,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[7.7495737E0,6.2250075E0,1.524566E0,2.1450396E0,4.079968E0,1.078177E0,1.0668626E0,1.0623941E0,3.0175738E0,1.7642405E0,1.2533333E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"11","size_leaf_vector":"0"}},{"base_weights":[7.5330045E-3,7.0384726E-2,-1.0434088E-1,-2.744488E-2,6.559661E-2,-6.664765E-2,1.6296057E-2,-1.11923374E-1,3.490192E-2,-5.893938E-2,1.1081644E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,true,false,false,false,true,false,false,false],"id":83,"left_children":[1,3,5,7,-1,-1,-1,9,-1,-1,-1],"loss_changes":[6.823415E-2,9.50895E-2,7.782743E-2,6.875997E-2,0E0,0E0,0E0,5.004857E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3,7,7],"right_children":[2,4,6,8,-1,-1,-1,10,-1,-1,-1],"split_conditions":[4.1914754E0,2.7519693E0,-8.000531E-2,3.9246242E0,6.559661E-2,-6.664765E-2,1.6296057E-2,1.3915041E-1,3.490192E-2,-5.893938E-2,1.1081644E-2],"split_indices":[110,158,387,131,0,0,0,439,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[7.659667E0,5.14181E0,2.517857E0,3.5888023E0,1.5530075E0,1.2152393E0,1.3026178E0,2.4001944E0,1.188608E0,1.3268692E0,1.0733252E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"11","size_leaf_vector":"0"}},{"base_weights":[5.5437726E-3,-5.0958794E-2,3.99763E-2,-1.4893697E-1,7.833306E-2,6.3577257E-3,-6.828566E-2,5.3801257E-2,-1.5503987E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,true,true,false,false,false,false],"id":84,"left_children":[1,3,-1,5,7,-1,-1,-1,-1],"loss_changes":[6.858371E-2,9.603622E-2,0E0,6.015628E-2,5.5578947E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,3,3,4,4],"right_children":[2,4,-1,6,8,-1,-1,-1,-1],"split_conditions":[5.5010567E0,6.1421316E-2,3.99763E-2,-3.2279758E0,-2.345273E-2,6.3577257E-3,-6.828566E-2,5.3801257E-2,-1.5503987E-2],"split_indices":[70,400,0,106,189,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[7.5633945E0,5.660167E0,1.9032274E0,3.133578E0,2.5265892E0,1.2219915E0,1.9115865E0,1.2084149E0,1.3181742E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[5.7673673E-3,1.0835938E-1,-6.3555606E-2,6.4139865E-2,-1.4990136E-2,-1.4571548E-1,3.720499E-2,1.5267186E-3,-7.0446484E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,false,true,false,false,false],"id":85,"left_children":[1,3,5,-1,-1,7,-1,-1,-1],"loss_changes":[6.7263246E-2,7.400237E-2,1.05858415E-1,0E0,0E0,6.060391E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,5,5],"right_children":[2,4,6,-1,-1,8,-1,-1,-1],"split_conditions":[-2.826052E-1,6.1446655E-1,6.43373E-1,6.4139865E-2,-1.4990136E-2,-1.645573E-2,3.720499E-2,1.5267186E-3,-7.0446484E-2],"split_indices":[363,159,364,0,0,189,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[7.4893875E0,2.7929513E0,4.6964364E0,1.466164E0,1.3267874E0,3.4210901E0,1.2753464E0,1.6208416E0,1.8002485E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[6.1116423E-3,7.523332E-2,-9.926468E-2,-2.7040234E-2,1.444548E-1,-5.9104033E-2,1.7857948E-2,7.069957E-2,-1.7272308E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,true,false,false,false,false],"id":86,"left_children":[1,3,5,-1,7,-1,-1,-1,-1],"loss_changes":[6.871048E-2,7.7653795E-2,7.011553E-2,0E0,9.399293E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,4,4],"right_children":[2,4,6,-1,8,-1,-1,-1,-1],"split_conditions":[3.0724077E0,-5.5330687E0,5.43046E-1,-2.7040234E-2,3.667283E-1,-5.9104033E-2,1.7857948E-2,7.069957E-2,-1.7272308E-2],"split_indices":[25,160,45,0,397,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[7.4080534E0,4.6463366E0,2.761717E0,1.2818753E0,3.3644614E0,1.5604355E0,1.2012815E0,2.2032607E0,1.1612009E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[2.6349556E-3,4.6297304E-2,-4.66302E-2,1.0582986E-1,-3.504832E-2,-1.6573487E-2,1.9366875E-1,-1.2135897E-4,7.807957E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false,false,false,false,false],"id":87,"left_children":[1,3,-1,5,-1,-1,7,-1,-1],"loss_changes":[6.466139E-2,8.320414E-2,0E0,9.3737066E-2,0E0,0E0,5.205977E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,3,3,6,6],"right_children":[2,4,-1,6,-1,-1,8,-1,-1],"split_conditions":[8.174833E0,5.296294E0,-4.66302E-2,-3.4296675E0,-3.504832E-2,-1.6573487E-2,-2.8782496E0,-1.2135897E-4,7.807957E-2],"split_indices":[147,117,0,59,0,0,138,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[7.3262873E0,6.2946615E0,1.031626E0,4.868965E0,1.4256963E0,1.8491489E0,3.0198164E0,1.0254393E0,1.9943771E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[-3.981059E-4,6.82583E-2,-1.0222523E-1,-2.4623588E-2,1.6880004E-1,2.12108E-2,-5.9537765E-2,7.410281E-2,-2.9597923E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,true,false,false,false,false],"id":88,"left_children":[1,3,5,-1,7,-1,-1,-1,-1],"loss_changes":[6.464591E-2,9.738879E-2,7.5976565E-2,0E0,5.3284287E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,4,4],"right_children":[2,4,6,-1,8,-1,-1,-1,-1],"split_conditions":[2.7524998E0,-9.9897707E-1,-1.0296428E0,-2.4623588E-2,2.4970295E-1,2.12108E-2,-5.9537765E-2,7.410281E-2,-2.9597923E-3],"split_indices":[132,161,122,0,260,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[7.2488174E0,4.526508E0,2.7223098E0,1.8876299E0,2.6388779E0,1.0681678E0,1.654142E0,1.5693778E0,1.0695001E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[8.06913E-4,-5.5498704E-2,3.5727426E-2,-1.2549913E-1,3.911426E-2,2.5390448E-2,-7.335047E-2,3.5727557E-2,-2.5628366E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,true,false,true,false,false,false],"id":89,"left_children":[1,3,-1,5,-1,7,-1,-1,-1],"loss_changes":[6.103062E-2,9.727079E-2,0E0,9.646658E-2,0E0,4.259498E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,3,3,5,5],"right_children":[2,4,-1,6,-1,8,-1,-1,-1],"split_conditions":[3.8640652E0,5.923786E0,3.5727426E-2,9.9676466E-1,3.911426E-2,9.3281007E-1,-7.335047E-2,3.5727557E-2,-2.5628366E-2],"split_indices":[156,161,0,121,0,159,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[7.1712503E0,5.218128E0,1.9531224E0,4.02659E0,1.1915381E0,2.1222782E0,1.9043117E0,1.1094959E0,1.0127823E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[1.8492481E-3,-6.840366E-2,9.446293E-2,-1.5735368E-1,2.7524369E-2,5.721452E-2,-1.8767918E-2,-7.460775E-2,8.410908E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,true,false,false,false,false,false],"id":90,"left_children":[1,3,5,7,-1,-1,-1,-1,-1],"loss_changes":[5.9112303E-2,8.825113E-2,7.103455E-2,6.961967E-2,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3],"right_children":[2,4,6,8,-1,-1,-1,-1,-1],"split_conditions":[1.7952664E0,2.6346316E0,2.3179899E-1,1.981225E0,2.7524369E-2,5.721452E-2,-1.8767918E-2,-7.460775E-2,8.410908E-3],"split_indices":[133,56,291,3,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[7.092163E0,4.18159E0,2.910573E0,2.6996458E0,1.4819442E0,1.6714396E0,1.2391334E0,1.5798316E0,1.1198142E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[1.1756576E-3,5.6761745E-2,-3.4110658E-2,-6.1761793E-2,1.5586998E-1,-5.169986E-2,2.6940452E-2,5.694497E-4,6.9634326E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,true,true,false,false,false,false],"id":91,"left_children":[1,3,-1,5,7,-1,-1,-1,-1],"loss_changes":[5.7636976E-2,8.091351E-2,0E0,7.3792875E-2,4.169043E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,3,3,4,4],"right_children":[2,4,-1,6,8,-1,-1,-1,-1],"split_conditions":[4.7324567E0,-1.5762246E-1,-3.4110658E-2,1.4821422E-1,3.3391E-2,-5.169986E-2,2.6940452E-2,5.694497E-4,6.9634326E-2],"split_indices":[134,78,0,271,79,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[7.0153728E0,5.068679E0,1.9466938E0,2.4798503E0,2.5888286E0,1.3545905E0,1.1252598E0,1.1968156E0,1.392013E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[2.3343845E-3,4.6499398E-2,-4.265245E-2,-3.4023497E-2,9.9675365E-2,-3.23325E-2,1.72613E-1,7.607319E-2,4.0738857E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,true,false,true,false,false],"id":92,"left_children":[1,3,-1,-1,5,-1,7,-1,-1],"loss_changes":[5.733475E-2,6.947443E-2,0E0,0E0,1.04333654E-1,0E0,5.5482313E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,4,4,6,6],"right_children":[2,4,-1,-1,6,-1,8,-1,-1],"split_conditions":[7.6444206E0,-8.643622E-1,-4.265245E-2,-3.4023497E-2,-6.314613E0,-3.23325E-2,1.12008624E-1,7.607319E-2,4.0738857E-3],"split_indices":[75,488,0,0,93,0,226,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[6.9476204E0,5.8407817E0,1.1068387E0,1.1748887E0,4.665893E0,1.089497E0,3.5763962E0,1.975946E0,1.6004503E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[-1.8275352E-3,3.575245E-2,-5.3188164E-2,3.144937E-2,-1.2600598E-1,-1.9903888E-1,1.1017406E-2,-8.173095E-2,-1.648766E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,true,true,false,false,false],"id":93,"left_children":[1,-1,3,-1,5,7,-1,-1,-1],"loss_changes":[5.5258986E-2,0E0,8.475143E-2,0E0,6.3290074E-2,2.10561E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,2,2,4,4,5,5],"right_children":[2,-1,4,-1,6,8,-1,-1,-1],"split_conditions":[-5.511676E0,3.575245E-2,-4.5544E0,3.144937E-2,3.4194243E-1,-1.16662666E-1,1.1017406E-2,-8.173095E-2,-1.648766E-2],"split_indices":[81,0,141,0,248,397,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[6.8716345E0,1.6541672E0,5.2174673E0,1.5071709E0,3.7102964E0,2.4069483E0,1.3033481E0,1.0044103E0,1.4025381E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[-4.027322E-3,4.8098285E-2,-3.676831E-2,5.457081E-2,-3.397029E-2,2.4779078E-2,-4.57663E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,true,false,false],"id":94,"left_children":[1,3,-1,-1,5,-1,-1],"loss_changes":[5.4160237E-2,7.364557E-2,0E0,0E0,7.6888114E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,4,4],"right_children":[2,4,-1,-1,6,-1,-1],"split_conditions":[6.333127E0,-3.7427216E0,-3.676831E-2,5.457081E-2,1.1759052E0,2.4779078E-2,-4.57663E-2],"split_indices":[110,145,0,0,139,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[6.8116302E0,5.143834E0,1.6677963E0,1.4930677E0,3.6507664E0,1.9940773E0,1.656689E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[-3.2367967E-3,3.8971305E-2,-5.153669E-2,-6.0704235E-2,3.3654653E-2,4.4773743E-2,-8.589956E-2,-4.8508346E-2,8.9908345E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false,false,false,false,false],"id":95,"left_children":[1,-1,3,-1,5,-1,7,-1,-1],"loss_changes":[5.658047E-2,0E0,8.914279E-2,0E0,8.1260376E-2,0E0,3.1800456E-2,0E0,0E0],"parents":[2147483647,0,0,2,2,4,4,6,6],"right_children":[2,-1,4,-1,6,-1,8,-1,-1],"split_conditions":[-7.211624E0,3.8971305E-2,-4.538929E0,-6.0704235E-2,7.3667336E-1,4.4773743E-2,2.1083379E-1,-4.8508346E-2,8.9908345E-3],"split_indices":[61,0,47,0,64,0,123,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[6.757415E0,1.3490772E0,5.4083376E0,1.4558661E0,3.9524717E0,1.8832737E0,2.069198E0,1.0118256E0,1.0573723E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[-4.5060283E-3,8.841913E-2,-7.2796635E-2,-2.4213986E-2,6.0372803E-2,2.7006026E-2,-1.4891584E-1,2.8121718E-3,-6.804761E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false,false,false,false,false],"id":96,"left_children":[1,3,5,-1,-1,-1,7,-1,-1],"loss_changes":[5.5224262E-2,8.700293E-2,7.5123176E-2,0E0,0E0,0E0,4.8620768E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,6,6],"right_children":[2,4,6,-1,-1,-1,8,-1,-1],"split_conditions":[-1.248637E0,-1.1932282E0,-2.719048E0,-2.4213986E-2,6.0372803E-2,2.7006026E-2,-4.645598E-1,2.8121718E-3,-6.804761E-2],"split_indices":[86,101,134,0,0,0,122,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[6.68804E0,2.7081828E0,3.979857E0,1.1975523E0,1.5106305E0,1.2097079E0,2.7701492E0,1.2038844E0,1.5662649E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[-3.5042732E-3,-5.8547772E-2,3.1597108E-2,2.3351097E-2,-1.4137098E-1,-6.607336E-2,4.104847E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false,false,false],"id":97,"left_children":[1,3,-1,-1,5,-1,-1],"loss_changes":[5.179527E-2,7.540493E-2,0E0,0E0,5.0736956E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,4,4],"right_children":[2,4,-1,-1,6,-1,-1],"split_conditions":[4.8019023E0,-2.3680513E-1,3.1597108E-2,2.3351097E-2,1.4847856E0,-6.607336E-2,4.104847E-3],"split_indices":[70,288,0,0,101,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[6.6171484E0,4.7013125E0,1.9158361E0,1.7990407E0,2.902272E0,1.6450305E0,1.2572415E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[-3.524258E-3,-5.515572E-2,3.3212546E-2,-1.1656132E-1,3.4075495E-2,-5.925544E-2,3.243203E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,true,false,false,false],"id":98,"left_children":[1,3,-1,5,-1,-1,-1],"loss_changes":[5.075326E-2,7.417552E-2,0E0,5.16864E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,3,3],"right_children":[2,4,-1,6,-1,-1,-1],"split_conditions":[5.077216E0,5.623352E0,3.3212546E-2,8.040628E-2,3.4075495E-2,-5.925544E-2,3.243203E-3],"split_indices":[4,45,0,228,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[6.5698595E0,4.8809214E0,1.6889383E0,3.8053596E0,1.0755619E0,1.9898839E0,1.8154757E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[-1.0722914E-3,-3.6605943E-2,4.7162313E-2,1.255896E-1,-7.481332E-2,6.785546E-2,-3.0466334E-3,-4.364781E-2,9.9389125E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,true,true,false,false,false,false],"id":99,"left_children":[1,-1,3,5,7,-1,-1,-1,-1],"loss_changes":[4.9642306E-2,0E0,6.774326E-2,5.8088236E-2,2.7812269E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,2,2,3,3,4,4],"right_children":[2,-1,4,6,8,-1,-1,-1,-1],"split_conditions":[-5.87751E0,-3.6605943E-2,1.6571867E0,-1.9067378E-1,5.1523484E-2,6.785546E-2,-3.0466334E-3,-4.364781E-2,9.9389125E-3],"split_indices":[135,0,147,238,391,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[6.522909E0,1.4235829E0,5.099326E0,3.0856814E0,2.0136447E0,1.3896325E0,1.696049E0,1.0066444E0,1.0070003E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[-1.3991267E-3,4.3586276E-2,-4.037258E-2,1.2379593E-1,-7.376729E-2,-1.7864093E-2,2.0198743E-1,1.2904792E-2,-4.4960905E-2,1.3964615E-2,7.7432E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,true,true,false,true,false,false,false,false],"id":100,"left_children":[1,3,-1,5,7,-1,9,-1,-1,-1,-1],"loss_changes":[5.0609488E-2,6.8791926E-2,0E0,7.135312E-2,3.4885414E-2,0E0,1.527752E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,3,3,4,4,6,6],"right_children":[2,4,-1,6,8,-1,10,-1,-1,-1,-1],"split_conditions":[6.3650026E0,2.0457262E-1,-4.037258E-2,-4.752125E0,-8.5826255E-2,-1.7864093E-2,-4.5291758E-1,1.2904792E-2,-4.4960905E-2,1.3964615E-2,7.7432E-2],"split_indices":[109,168,0,48,180,0,62,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0,0,0],"sum_hessian":[6.4685335E0,5.3381E0,1.1304334E0,3.1382499E0,2.19985E0,1.0095385E0,2.1287115E0,1.0394775E0,1.1603726E0,1.0499701E0,1.0787413E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"11","size_leaf_vector":"0"}},{"base_weights":[-7.0840766E-4,-6.3452E-2,9.352092E-2,2.0370176E-2,-1.4617035E-1,5.263931E-2,-1.1574003E-2,-6.902745E-2,1.2252114E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,true,false,false,false,false],"id":101,"left_children":[1,3,5,-1,7,-1,-1,-1,-1],"loss_changes":[4.9707297E-2,6.468807E-2,4.3179527E-2,0E0,4.393585E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,4,4],"right_children":[2,4,6,-1,8,-1,-1,-1,-1],"split_conditions":[3.959839E-1,-1.8640366E0,-8.187554E-1,2.0370176E-2,3.8938727E-2,5.263931E-2,-1.1574003E-2,-6.902745E-2,1.2252114E-3],"split_indices":[177,117,63,0,490,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[6.4037714E0,4.0401897E0,2.3635814E0,1.6303737E0,2.409816E0,1.2561314E0,1.10745E0,1.2052852E0,1.2045308E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[-1.4451216E-3,9.9236965E-2,-6.1111365E-2,-1.4606309E-2,5.9509736E-2,-1.4362097E-1,3.808191E-2,5.371365E-3,-6.2441416E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,false,false,false,false,false,false],"id":102,"left_children":[1,3,5,-1,-1,7,-1,-1,-1],"loss_changes":[5.021687E-2,5.534959E-2,9.9438466E-2,0E0,0E0,4.652486E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,5,5],"right_children":[2,4,6,-1,-1,8,-1,-1,-1],"split_conditions":[-3.1927748E0,-3.8271856E-1,6.3103753E-1,-1.4606309E-2,5.9509736E-2,-3.0631822E-1,3.808191E-2,5.371365E-3,-6.2441416E-2],"split_indices":[106,48,296,0,0,173,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[6.349758E0,2.11599E0,4.233768E0,1.0531999E0,1.06279E0,3.106862E0,1.1269062E0,1.0929714E0,2.0138905E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[2.5003422E-3,-9.888579E-2,6.0538467E-2,7.6117367E-3,-5.2415404E-2,1.2127053E-1,-2.7357483E-2,-7.944885E-3,6.3926876E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false,true,false,false,false],"id":103,"left_children":[1,3,5,-1,-1,7,-1,-1,-1],"loss_changes":[4.8840545E-2,3.2993834E-2,5.9453603E-2,0E0,0E0,6.2348787E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,5,5],"right_children":[2,4,6,-1,-1,8,-1,-1,-1],"split_conditions":[-3.208375E0,-6.456974E-2,5.3240094E0,7.6117367E-3,-5.2415404E-2,-2.0247204E0,-2.7357483E-2,-7.944885E-3,6.3926876E-2],"split_indices":[153,302,141,0,0,55,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[6.2828426E0,2.0310385E0,4.2518044E0,1.0219274E0,1.0091112E0,3.1797955E0,1.0720088E0,1.4914143E0,1.6883812E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[-3.9764884E-4,3.4203455E-2,-5.2307744E-2,-1.3987796E-1,4.575932E-2,-5.0126336E-4,-6.716875E-2,4.5193E-2,-1.8899554E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,true,true,false,false,false,false],"id":104,"left_children":[1,-1,3,5,7,-1,-1,-1,-1],"loss_changes":[4.889253E-2,0E0,5.4988254E-2,3.7194252E-2,4.852979E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,2,2,3,3,4,4],"right_children":[2,-1,4,6,8,-1,-1,-1,-1],"split_conditions":[-6.518854E-1,3.4203455E-2,-8.098646E-3,2.5717798E-1,5.153173E-1,-5.0126336E-4,-6.716875E-2,4.5193E-2,-1.8899554E-2],"split_indices":[212,0,182,275,14,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[6.228307E0,1.5705371E0,4.6577697E0,2.2353425E0,2.4224272E0,1.2307234E0,1.0046191E0,1.0371209E0,1.3853062E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[-1.3518782E-3,-7.292545E-2,8.020931E-2,1.5083259E-2,-4.7815125E-2,-1.2393518E-2,5.1278982E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false,false,false],"id":105,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[4.7748376E-2,5.4312546E-2,4.9294673E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[2.0019817E-1,-2.0152843E-1,-6.896371E-2,1.5083259E-2,-4.7815125E-2,-1.2393518E-2,5.1278982E-2],"split_indices":[19,80,400,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[6.1774187E0,3.3465517E0,2.830867E0,1.5525898E0,1.7939619E0,1.4428222E0,1.3880448E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[-8.488126E-5,-5.090567E-2,3.5171695E-2,-1.21758066E-1,3.3672247E-2,1.4739885E-2,-1.9128777E-1,-1.5474119E-2,-7.5694375E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false,false,true,false,false],"id":106,"left_children":[1,3,-1,5,-1,-1,7,-1,-1],"loss_changes":[4.852359E-2,7.929159E-2,0E0,6.2129848E-2,0E0,0E0,1.5622571E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,3,3,6,6],"right_children":[2,4,-1,6,-1,-1,8,-1,-1],"split_conditions":[5.923786E0,3.8640652E0,3.5171695E-2,-3.0461774E0,3.3672247E-2,1.4739885E-2,9.9676466E-1,-1.5474119E-2,-7.5694375E-2],"split_indices":[161,156,0,125,0,0,121,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[6.1371956E0,4.6772685E0,1.4599272E0,3.4379153E0,1.2393531E0,1.0790833E0,2.3588321E0,1.2781041E0,1.080728E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[-1.2869949E-3,5.8054306E-2,-9.8977454E-2,-2.9283052E-2,1.2689385E-1,-5.580381E-2,1.0857177E-2,6.1261624E-2,-7.985924E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,true,false,false,false,false],"id":107,"left_children":[1,3,5,-1,7,-1,-1,-1,-1],"loss_changes":[4.6798553E-2,6.634911E-2,4.2563662E-2,0E0,5.203656E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,4,4],"right_children":[2,4,6,-1,8,-1,-1,-1,-1],"split_conditions":[4.1914754E0,-4.649165E0,-3.511431E-2,-2.9283052E-2,3.3159822E-1,-5.580381E-2,1.0857177E-2,6.1261624E-2,-7.985924E-3],"split_indices":[110,90,387,0,333,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[6.081582E0,4.0357995E0,2.0457826E0,1.109343E0,2.9264565E0,1.0156456E0,1.030137E0,1.7266707E0,1.1997857E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[-3.8436824E-3,-6.964319E-2,8.491437E-2,-5.1193897E-2,1.320834E-2,5.276597E-2,-1.1051586E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,false,false,false,false],"id":108,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[4.6947345E-2,6.014553E-2,4.460097E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[1.86212E0,-1.1873546E0,-1.548356E0,-5.1193897E-2,1.320834E-2,5.276597E-2,-1.1051586E-2],"split_indices":[154,108,57,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[6.0260243E0,3.5842507E0,2.4417737E0,1.6324744E0,1.9517763E0,1.1430695E0,1.2987041E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[-2.1798722E-3,-9.0556294E-2,6.6462375E-2,1.946237E-2,-5.919708E-2,-2.9523196E-2,1.4211361E-1,1.6387174E-3,6.0494434E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false,false,false,false,false],"id":109,"left_children":[1,3,5,-1,-1,-1,7,-1,-1],"loss_changes":[4.83944E-2,7.0501946E-2,7.02178E-2,0E0,0E0,0E0,2.7368762E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,6,6],"right_children":[2,4,6,-1,-1,-1,8,-1,-1],"split_conditions":[-1.4493701E0,-1.8379398E-1,-5.318883E0,1.946237E-2,-5.919708E-2,-2.9523196E-2,-1.8817723E0,1.6387174E-3,6.0494434E-2],"split_indices":[5,261,160,0,0,0,76,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[5.9853816E0,2.4770048E0,3.508377E0,1.1684134E0,1.3085914E0,1.0088531E0,2.4995239E0,1.0898081E0,1.4097158E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[-2.563621E-3,6.4668335E-2,-9.170216E-2,4.617552E-2,-1.6473256E-2,1.3506646E-2,-5.208214E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,false,false,false],"id":110,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[4.7445487E-2,5.6855097E-2,4.5801006E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[2.1612035E-1,-3.6670685E-2,-1.194378E-1,4.617552E-2,-1.6473256E-2,1.3506646E-2,-5.208214E-2],"split_indices":[258,164,428,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.92734E0,3.5353515E0,2.3919885E0,1.8599658E0,1.6753857E0,1.0648094E0,1.3271791E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[-3.1055955E-3,-7.28988E-2,7.960327E-2,2.1289274E-2,-4.584102E-2,4.891925E-2,-1.2270066E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,false,false,false],"id":111,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[4.5560446E-2,5.978281E-2,4.317798E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[1.5412903E0,-2.4530811E0,-4.3729258E-1,2.1289274E-2,-4.584102E-2,4.891925E-2,-1.2270066E-2],"split_indices":[137,61,98,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.8873706E0,3.2573185E0,2.6300519E0,1.203101E0,2.0542176E0,1.3451848E0,1.284867E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[-4.2028175E-4,-3.3781692E-2,5.101643E-2,-3.38844E-2,5.1098924E-2,2.0581849E-2,-4.083857E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false,false,false],"id":112,"left_children":[1,-1,3,5,-1,-1,-1],"loss_changes":[4.525828E-2,0E0,6.0301803E-2,5.1037688E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,2,2,3,3],"right_children":[2,-1,4,6,-1,-1,-1],"split_conditions":[-5.2224083E0,-3.3781692E-2,2.853623E0,8.2212687E-1,5.1098924E-2,2.0581849E-2,-4.083857E-2],"split_indices":[73,0,52,107,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.8475723E0,1.4644173E0,4.383155E0,2.979214E0,1.403941E0,1.6521147E0,1.3270992E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[-7.0803205E-4,6.0862362E-2,-3.028845E-2,-2.2707094E-2,1.5002933E-1,6.114414E-2,5.77715E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,true,false,false],"id":113,"left_children":[1,3,-1,-1,5,-1,-1],"loss_changes":[4.8153747E-2,7.0289515E-2,0E0,0E0,1.9028716E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,4,4],"right_children":[2,4,-1,-1,6,-1,-1],"split_conditions":[3.4428743E-1,-2.7586539E0,-3.028845E-2,-2.2707094E-2,5.852413E-1,6.114414E-2,5.77715E-3],"split_indices":[410,71,0,0,46,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.8056707E0,3.8401613E0,1.9655093E0,1.5767002E0,2.263461E0,1.2080631E0,1.055398E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[1.6478588E-3,6.3221425E-2,-9.441999E-2,1.3096847E-1,-2.096552E-2,-5.530634E-2,1.1914198E-2,-9.853013E-3,6.6231586E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,true,false,false,false,false,false],"id":114,"left_children":[1,3,5,7,-1,-1,-1,-1,-1],"loss_changes":[4.6009943E-2,5.1752336E-2,4.4053957E-2,5.8424115E-2,0E0,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2,3,3],"right_children":[2,4,6,8,-1,-1,-1,-1,-1],"split_conditions":[2.2912774E0,3.8289309E0,-7.277336E-1,-7.366648E-1,-2.096552E-2,-5.530634E-2,1.1914198E-2,-9.853013E-3,6.6231586E-2],"split_indices":[81,132,92,11,0,0,0,0,0],"split_type":[0,0,0,0,0,0,0,0,0],"sum_hessian":[5.769053E0,3.7240653E0,2.0449877E0,2.4785986E0,1.2454666E0,1.0000598E0,1.0449278E0,1.1022476E0,1.376351E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"9","size_leaf_vector":"0"}},{"base_weights":[5.035352E-5,-4.841779E-2,3.556601E-2,3.2134593E-2,-1.1368565E-1,1.7427979E-2,-5.7645664E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false,false,false],"id":115,"left_children":[1,3,-1,-1,5,-1,-1],"loss_changes":[4.4307888E-2,6.776328E-2,0E0,0E0,6.969145E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,4,4],"right_children":[2,4,-1,-1,6,-1,-1],"split_conditions":[4.732904E0,-6.235343E-1,3.556601E-2,3.2134593E-2,-4.3129416E0,1.7427979E-2,-5.7645664E-2],"split_indices":[6,403,0,0,136,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.7148933E0,4.4757266E0,1.2391665E0,1.1334805E0,3.342246E0,1.1294025E0,2.2128437E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[4.059812E-3,-2.8726444E-2,6.163376E-2,-1.6975038E-2,1.4322802E-1,1.1321648E-3,6.277478E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,true,false,false],"id":116,"left_children":[1,-1,3,-1,5,-1,-1],"loss_changes":[4.4213384E-2,0E0,5.475519E-2,0E0,2.8162338E-2,0E0,0E0],"parents":[2147483647,0,0,2,2,4,4],"right_children":[2,-1,4,-1,6,-1,-1],"split_conditions":[-3.5643492E0,-2.8726444E-2,-2.057909E-1,-1.6975038E-2,-1.6846621E-1,1.1321648E-3,6.277478E-2],"split_indices":[154,0,319,0,362,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.6666975E0,1.8303314E0,3.8363662E0,1.6917762E0,2.14459E0,1.0287536E0,1.1158363E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[5.705035E-3,5.8644373E-2,-3.1195672E-2,1.3899341E-1,-1.3583544E-2,4.74356E-3,5.7667386E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,true,false,false,false],"id":117,"left_children":[1,3,-1,5,-1,-1,-1],"loss_changes":[4.461174E-2,4.898555E-2,0E0,1.7478637E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,3,3],"right_children":[2,4,-1,6,-1,-1,-1],"split_conditions":[5.805045E-1,-1.43278E-2,-3.1195672E-2,-1.740381E0,-1.3583544E-2,4.74356E-3,5.7667386E-2],"split_indices":[425,176,0,112,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.632325E0,4.112769E0,1.5195562E0,2.1291373E0,1.983632E0,1.0338233E0,1.095314E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[6.7564417E-3,9.2019975E-2,-5.680537E-2,4.92394E-2,-9.832606E-3,2.3173597E-2,-4.653482E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false,false,false],"id":118,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[4.100909E-2,3.4190446E-2,6.997981E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[-1.7640333E0,1.5379213E-1,-1.2528253E0,4.92394E-2,-9.832606E-3,2.3173597E-2,-4.653482E-2],"split_indices":[64,302,72,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.6024227E0,2.2015197E0,3.400903E0,1.1955074E0,1.0060123E0,1.5295594E0,1.8713436E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[7.20302E-3,2.9676847E-2,-5.152614E-2,-1.2184916E-1,2.940526E-2,2.9081178E-3,-5.4126587E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,false,false,false],"id":119,"left_children":[1,-1,3,5,-1,-1,-1],"loss_changes":[4.0452465E-2,0E0,6.0938243E-2,2.9598825E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,2,2,3,3],"right_children":[2,-1,4,6,-1,-1,-1],"split_conditions":[-3.947982E0,2.9676847E-2,6.5138865E-1,-2.2192607E0,2.940526E-2,2.9081178E-3,-5.4126587E-2],"split_indices":[162,0,404,149,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.563543E0,1.9046229E0,3.6589198E0,2.614597E0,1.0443227E0,1.0626335E0,1.5519636E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[6.2083537E-3,-6.911396E-2,7.525784E-2,-6.1773796E-2,2.4289263E-2,4.4308428E-2,-1.938082E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,false,false,false],"id":120,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[3.9156403E-2,9.1479495E-2,4.889919E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[-5.077629E-1,-1.9369357E0,3.4428743E-1,-6.1773796E-2,2.4289263E-2,4.4308428E-2,-1.938082E-2],"split_indices":[47,10,410,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.5285997E0,2.643747E0,2.8848526E0,1.188434E0,1.4553131E0,1.8636254E0,1.0212272E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[2.3807958E-3,3.602094E-2,-4.3429542E-2,2.8798157E-2,-1.1004187E-1,8.944857E-3,-5.671769E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,true,false,false],"id":121,"left_children":[1,-1,3,-1,5,-1,-1],"loss_changes":[4.015338E-2,0E0,6.0706735E-2,0E0,5.1162034E-2,0E0,0E0],"parents":[2147483647,0,0,2,2,4,4],"right_children":[2,-1,4,-1,6,-1,-1],"split_conditions":[-9.139375E-1,3.602094E-2,-3.6159916E0,2.8798157E-2,-1.8246196E-1,8.944857E-3,-5.671769E-2],"split_indices":[285,0,159,0,334,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.480502E0,1.0813758E0,4.3991265E0,1.2796541E0,3.1194723E0,1.3509644E0,1.7685078E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[4.2404756E-3,-4.5189247E-2,3.2649145E-2,-1.0965811E-1,2.5477782E-2,1.06159905E-2,-5.5033457E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,true,false,false,false],"id":122,"left_children":[1,3,-1,5,-1,-1,-1],"loss_changes":[3.821107E-2,5.19233E-2,0E0,4.7515474E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,3,3],"right_children":[2,4,-1,6,-1,-1,-1],"split_conditions":[5.5010567E0,3.2638206E0,3.2649145E-2,-2.1737266E-1,2.5477782E-2,1.06159905E-2,-5.5033457E-2],"split_indices":[70,65,0,238,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.4398527E0,4.0797048E0,1.3601481E0,2.8331645E0,1.2465403E0,1.130779E0,1.7023854E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[5.048024E-3,-4.44465E-2,3.127699E-2,2.4970189E-2,-1.1016574E-1,-5.530982E-2,5.217437E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false,false,false],"id":123,"left_children":[1,3,-1,-1,5,-1,-1],"loss_changes":[3.6074795E-2,5.0898124E-2,0E0,0E0,3.835371E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,4,4],"right_children":[2,4,-1,-1,6,-1,-1],"split_conditions":[4.6290092E0,-4.5676097E-1,3.127699E-2,2.4970189E-2,-1.117568E0,-5.530982E-2,5.217437E-3],"split_indices":[46,249,0,0,108,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.403085E0,3.9729805E0,1.4301049E0,1.2594981E0,2.7134824E0,1.4339755E0,1.2795069E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[4.4454527E-3,-4.354966E-2,3.3101827E-2,-1.0242044E-1,2.6988534E-2,1.0345506E-2,-5.1006604E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,true,false,false,false],"id":124,"left_children":[1,3,-1,5,-1,-1,-1],"loss_changes":[3.717737E-2,4.929917E-2,0E0,4.2947542E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,3,3],"right_children":[2,4,-1,6,-1,-1,-1],"split_conditions":[5.923786E0,3.8640652E0,3.3101827E-2,-2.1055868E0,2.6988534E-2,1.0345506E-2,-5.1006604E-2],"split_indices":[161,156,0,125,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.3694654E0,4.0999546E0,1.2695107E0,3.0069401E0,1.0930144E0,1.1559038E0,1.8510363E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[6.135105E-3,7.352565E-2,-6.855079E-2,-1.7363591E-2,4.7728997E-2,1.4716305E-2,-4.8217822E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false,false,false],"id":125,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[3.6923505E-2,5.207725E-2,4.6066128E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[4.0255515E-3,-3.9565835E-2,1.0035095E0,-1.7363591E-2,4.7728997E-2,1.4716305E-2,-4.8217822E-2],"split_indices":[260,366,106,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.334687E0,2.8124728E0,2.5222144E0,1.2368157E0,1.5756571E0,1.3137869E0,1.2084275E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[3.8275982E-3,5.3247698E-2,-2.9464545E-2,4.8804954E-2,-4.3257378E-2,-4.0364128E-2,1.6817512E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false,false,false],"id":126,"left_children":[1,3,-1,-1,5,-1,-1],"loss_changes":[3.699253E-2,5.869989E-2,0E0,0E0,3.8272686E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,4,4],"right_children":[2,4,-1,-1,6,-1,-1],"split_conditions":[5.099218E0,-2.4530811E0,-2.9464545E-2,4.8804954E-2,-5.077629E-1,-4.0364128E-2,1.6817512E-2],"split_indices":[23,61,0,0,47,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.2984405E0,3.8917956E0,1.4066447E0,1.5023749E0,2.3894207E0,1.0601784E0,1.3292423E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[2.0481653E-3,3.3128586E-2,-4.575797E-2,-1.1561626E-1,2.1701787E-2,-5.620795E-2,6.052864E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false,false,false],"id":127,"left_children":[1,-1,3,5,-1,-1,-1],"loss_changes":[3.7499897E-2,0E0,5.004785E-2,3.845815E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,2,2,3,3],"right_children":[2,-1,4,6,-1,-1,-1],"split_conditions":[-6.0323806E0,3.3128586E-2,2.7421346E0,1.468729E0,2.1701787E-2,-5.620795E-2,6.052864E-3],"split_indices":[81,0,78,137,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.2623696E0,1.209778E0,4.052592E0,2.5595465E0,1.4930453E0,1.4262574E0,1.1332891E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[2.111767E-3,-4.9697246E-2,3.0352753E-2,2.0889908E-2,-1.11029886E-1,-5.440826E-2,5.9226267E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,true,false,false],"id":128,"left_children":[1,3,-1,-1,5,-1,-1],"loss_changes":[3.6994882E-2,4.2570524E-2,0E0,0E0,3.604339E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,4,4],"right_children":[2,4,-1,-1,6,-1,-1],"split_conditions":[5.0954676E-1,-3.947982E0,3.0352753E-2,2.0889908E-2,1.5133151E-1,-5.440826E-2,5.9226267E-3],"split_indices":[487,162,0,0,312,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.2284026E0,3.7602046E0,1.468198E0,1.2305951E0,2.5296094E0,1.3933815E0,1.136228E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[1.8572592E-3,6.746345E-2,-7.623643E-2,-1.2018892E-2,4.7254987E-2,-5.0149433E-2,1.1803986E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false,false,false],"id":129,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[3.6895495E-2,4.406055E-2,4.124841E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[1.3957454E-1,3.410101E-3,-2.920134E-1,-1.2018892E-2,4.7254987E-2,-5.0149433E-2,1.1803986E-2],"split_indices":[226,31,11,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.1974764E0,2.8985436E0,2.2989328E0,1.5741163E0,1.3244272E0,1.0369207E0,1.2620121E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[7.9912605E-4,-4.468749E-2,3.4224004E-2,2.0576553E-2,-1.1359931E-1,-5.811042E-2,4.2859808E-4],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,true,false,false],"id":130,"left_children":[1,3,-1,-1,5,-1,-1],"loss_changes":[3.6861368E-2,4.773368E-2,0E0,0E0,3.305884E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,4,4],"right_children":[2,4,-1,-1,6,-1,-1],"split_conditions":[8.661072E-1,-2.826052E-1,3.4224004E-2,2.0576553E-2,3.453826E-1,-5.811042E-2,4.2859808E-4],"split_indices":[228,363,0,0,363,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.1643457E0,4.116806E0,1.0475396E0,1.5589402E0,2.5578659E0,1.1046617E0,1.4532042E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[1.4232967E-5,-6.9811195E-2,7.032718E-2,-4.560597E-2,1.686575E-2,-1.1557128E-2,5.2065372E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false,false,false],"id":131,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[3.501854E-2,4.5625836E-2,4.6616502E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[-5.783028E-3,2.3680219E-1,1.7952664E0,-4.560597E-2,1.686575E-2,-1.1557128E-2,5.2065372E-2],"split_indices":[264,247,133,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.1326222E0,2.5788193E0,2.553803E0,1.4359457E0,1.1428735E0,1.5481043E0,1.0056987E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[1.9553897E-3,8.008634E-2,-5.9751213E-2,-1.8870333E-2,5.647032E-2,-4.2763956E-2,2.0628959E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,false,false,false,false],"id":132,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[3.4186106E-2,6.006074E-2,5.2150317E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[-2.106997E0,-9.140148E-3,1.204833E0,-1.8870333E-2,5.647032E-2,-4.2763956E-2,2.0628959E-2],"split_indices":[22,298,151,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.098249E0,2.1182854E0,2.9799635E0,1.0923799E0,1.0259055E0,1.7459497E0,1.2340138E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[3.9706076E-3,-5.6300554E-2,8.117294E-2,-4.2647712E-2,8.186499E-3,3.344494E-2,2.903491E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false,false,false],"id":133,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[3.276232E-2,3.2536197E-2,5.564792E-3,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[1.86212E0,-2.2397411E-1,5.969081E-1,-4.2647712E-2,8.186499E-3,3.344494E-2,2.903491E-3],"split_indices":[154,436,4,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.0588417E0,2.9929855E0,2.0658562E0,1.1307946E0,1.8621908E0,1.0579965E0,1.0078597E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[3.0782286E-3,6.798721E-2,-6.8513565E-2,-1.6904324E-2,5.1010866E-2,1.7693197E-2,-4.458488E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,false,false,false],"id":134,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[3.272396E-2,5.61419E-2,4.318115E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[2.1239257E-1,2.7550554E-1,-1.8599157E0,-1.6904324E-2,5.1010866E-2,1.7693197E-2,-4.458488E-2],"split_indices":[147,16,34,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[5.039641E0,2.6695929E0,2.370048E0,1.4052699E0,1.264323E0,1.0162772E0,1.3537709E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[3.3923592E-3,5.177946E-2,-2.7527463E-2,3.9502177E-2,-1.3468737E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":135,"left_children":[1,3,-1,-1,-1],"loss_changes":[3.2394458E-2,4.1590706E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[4.7324567E0,4.308008E-1,-2.7527463E-2,3.9502177E-2,-1.3468737E-2],"split_indices":[134,69,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[5.004107E0,3.6193616E0,1.3847455E0,1.7834544E0,1.8359072E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[2.1620041E-3,-2.926464E-2,4.851518E-2,-1.3643214E-2,3.880115E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,false],"id":136,"left_children":[1,-1,3,-1,-1],"loss_changes":[3.2391615E-2,0E0,4.1989304E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-5.332678E0,-2.926464E-2,-6.960285E-2,-1.3643214E-2,3.880115E-2],"split_indices":[73,0,5,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.9843245E0,1.231266E0,3.7530584E0,1.9373345E0,1.8157239E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[8.559025E-4,-6.330637E-2,7.303314E-2,1.6965479E-2,-3.989857E-2,-8.914603E-3,4.162091E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false,false,false],"id":137,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[3.224465E-2,3.8369127E-2,2.6206072E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[2.0019817E-1,-5.343344E-1,-1.462903E-1,1.6965479E-2,-3.989857E-2,-8.914603E-3,4.162091E-2],"split_indices":[19,14,400,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[4.9643383E0,2.6931524E0,2.2711859E0,1.0594742E0,1.6336782E0,1.099497E0,1.1716889E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[2.247017E-3,6.6895925E-2,-6.86973E-2,4.2097144E-2,-1.611147E-2,1.1946238E-2,-4.10152E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,false,false,false],"id":138,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[3.1831443E-2,3.953661E-2,2.998482E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[1.1673189E-1,3.0193767E-1,-1.2935688E-1,4.2097144E-2,-1.611147E-2,1.1946238E-2,-4.10152E-2],"split_indices":[280,448,290,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[4.9383054E0,2.6136491E0,2.3246562E0,1.5228946E0,1.0907545E0,1.0554231E0,1.2692331E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[1.0647913E-3,-7.354348E-2,6.1034214E-2,-4.4878043E-2,1.2267516E-2,-1.651032E-2,4.19879E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,true,false,false,false,false],"id":139,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[3.0936334E-2,3.2987833E-2,4.272913E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[-1.8478053E0,1.4252451E-1,-1.4768566E0,-4.4878043E-2,1.2267516E-2,-1.651032E-2,4.19879E-2],"split_indices":[94,231,38,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[4.91113E0,2.0875938E0,2.8235362E0,1.0695633E0,1.0180305E0,1.2653714E0,1.5581647E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[-4.6549828E-4,-5.1656593E-2,2.6155386E-2,3.595756E-2,-4.7100224E-2,3.682565E-2,-2.2122843E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false,false,false],"id":140,"left_children":[1,3,-1,5,-1,-1,-1],"loss_changes":[3.089691E-2,4.5715053E-2,0E0,3.936646E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,3,3],"right_children":[2,4,-1,6,-1,-1,-1],"split_conditions":[2.0623732E0,2.7524998E0,2.6155386E-2,1.6349456E-1,-4.7100224E-2,3.682565E-2,-2.2122843E-2],"split_indices":[112,132,0,269,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[4.882283E0,3.341416E0,1.5408674E0,2.1838148E0,1.157601E0,1.1527699E0,1.0310448E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[-1.419759E-3,2.9094668E-2,-4.8069067E-2,-4.379399E-2,3.556846E-2,-1.9727664E-2,3.6235277E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,true,false,false],"id":141,"left_children":[1,-1,3,-1,5,-1,-1],"loss_changes":[3.151951E-2,0E0,4.3171287E-2,0E0,3.6482796E-2,0E0,0E0],"parents":[2147483647,0,0,2,2,4,4],"right_children":[2,-1,4,-1,6,-1,-1],"split_conditions":[-5.9482584E0,2.9094668E-2,-2.7578666E0,-4.379399E-2,-8.406782E-2,-1.9727664E-2,3.6235277E-2],"split_indices":[63,0,103,0,16,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[4.850853E0,1.2130607E0,3.637792E0,1.332501E0,2.305291E0,1.157395E0,1.1478959E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[-7.031202E-4,5.7490114E-2,-2.65961E-2,4.664871E-2,-1.1059254E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":142,"left_children":[1,3,-1,-1,-1],"loss_changes":[3.488842E-2,4.3735072E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[2.2912774E0,-4.645598E-1,-2.65961E-2,4.664871E-2,-1.1059254E-2],"split_indices":[81,122,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.8208227E0,3.109639E0,1.7111837E0,1.2074575E0,1.9021814E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.1019487E-3,5.77863E-2,-2.4762405E-2,4.782788E-2,-1.1270937E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":143,"left_children":[1,3,-1,-1,-1],"loss_changes":[3.2266952E-2,4.519434E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.0724077E0,-3.4709337E-1,-2.4762405E-2,4.782788E-2,-1.1270937E-2],"split_indices":[25,425,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.7995462E0,3.0450654E0,1.7544808E0,1.1487304E0,1.896335E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[2.8655059E-3,2.4315434E-2,-5.775388E-2,1.368829E-2,-4.362055E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":144,"left_children":[1,-1,3,-1,-1],"loss_changes":[3.2069746E-2,0E0,4.1419685E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-1.248637E0,2.4315434E-2,1.2084732E0,1.368829E-2,-4.362055E-2],"split_indices":[86,0,134,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.778716E0,1.939773E0,2.8389432E0,1.5225288E0,1.3164145E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[9.988432E-4,6.833658E-2,-6.824694E-2,-1.5271301E-2,4.307782E-2,-4.5632433E-2,1.846047E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false,false,false],"id":145,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[3.151682E-2,3.7993528E-2,4.5706667E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[3.3998775E-1,-2.1762913E-1,3.497243E-1,-1.5271301E-2,4.307782E-2,-4.5632433E-2,1.846047E-2],"split_indices":[159,376,129,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[4.758931E0,2.4193625E0,2.3395686E0,1.0613204E0,1.3580421E0,1.3167164E0,1.0228522E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[8.062004E-4,3.0688135E-2,-4.4858005E-2,-1.08587965E-1,1.9398443E-2,6.621079E-3,-5.5769842E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,true,false,false,false],"id":146,"left_children":[1,-1,3,5,-1,-1,-1],"loss_changes":[3.113349E-2,0E0,3.944573E-2,3.7537377E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,2,2,3,3],"right_children":[2,-1,4,6,-1,-1,-1],"split_conditions":[-5.6195126E0,3.0688135E-2,2.042634E0,-2.5591766E-2,1.9398443E-2,6.621079E-3,-5.5769842E-2],"split_indices":[145,0,16,176,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[4.7278366E0,1.0823065E0,3.64553E0,2.3098953E0,1.3356347E0,1.1243113E0,1.185584E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[6.499715E-4,-2.6017861E-2,5.3027924E-2,-2.0299643E-2,1.1932365E-1,-3.207457E-4,5.3843092E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false,false,false],"id":147,"left_children":[1,-1,3,-1,5,-1,-1],"loss_changes":[3.0679066E-2,0E0,4.154745E-2,0E0,2.2367533E-2,0E0,0E0],"parents":[2147483647,0,0,2,2,4,4],"right_children":[2,-1,4,-1,6,-1,-1],"split_conditions":[-3.5643492E0,-2.6017861E-2,-3.5342927E0,-2.0299643E-2,2.8254485E-1,-3.207457E-4,5.3843092E-2],"split_indices":[154,0,7,0,78,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[4.698606E0,1.5151975E0,3.1834087E0,1.1213306E0,2.062078E0,1.0142837E0,1.0477943E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[1.7975137E-3,-4.5903657E-2,2.924871E-2,1.9715697E-2,-1.17427714E-1,-4.8050016E-2,-5.645808E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false,false,false],"id":148,"left_children":[1,3,-1,-1,5,-1,-1],"loss_changes":[3.03716E-2,4.337011E-2,0E0,0E0,9.9685155E-3,0E0,0E0],"parents":[2147483647,0,0,1,1,4,4],"right_children":[2,4,-1,-1,6,-1,-1],"split_conditions":[5.077216E0,-1.8928552E0,2.924871E-2,1.9715697E-2,-2.7825575E-2,-4.8050016E-2,-5.645808E-3],"split_indices":[4,164,0,0,351,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[4.6728735E0,3.4657063E0,1.2071675E0,1.3851544E0,2.0805519E0,1.0159471E0,1.0646048E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[3.8420263E-4,2.4407772E-2,-5.5991784E-2,1.7438352E-2,-3.8306344E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":149,"left_children":[1,-1,3,-1,-1],"loss_changes":[3.0348636E-2,0E0,3.9523005E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-2.64045E-1,2.4407772E-2,-1.9617252E-1,1.7438352E-2,-3.8306344E-2],"split_indices":[390,0,421,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.6501465E0,1.726768E0,2.9233785E0,1.2009904E0,1.722388E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[8.708613E-4,4.6346914E-2,-2.8838772E-2,-1.6699934E-2,1.1145157E-1,-1.3568498E-3,5.1485416E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,true,false,false],"id":150,"left_children":[1,3,-1,-1,5,-1,-1],"loss_changes":[2.9297365E-2,3.61495E-2,0E0,0E0,2.230534E-2,0E0,0E0],"parents":[2147483647,0,0,1,1,4,4],"right_children":[2,4,-1,-1,6,-1,-1],"split_conditions":[6.333127E0,-1.3897004E0,-2.8838772E-2,-1.6699934E-2,7.228753E-2,-1.3568498E-3,5.1485416E-2],"split_indices":[110,36,0,0,228,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[4.6316133E0,3.5087986E0,1.1228148E0,1.4234095E0,2.0853891E0,1.0282346E0,1.0571545E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[1.4908428E-3,5.7559606E-2,-2.336954E-2,4.0374123E-2,-1.9229056E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":151,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.9445374E-2,4.481031E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[2.0011115E0,2.7233639E0,-2.336954E-2,4.0374123E-2,-1.9229056E-2],"split_indices":[117,125,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.607753E0,2.861668E0,1.7460848E0,1.6872412E0,1.1744269E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[6.838897E-4,2.9548613E-2,-4.405916E-2,2.3654656E-2,-1.1002575E-1,-5.0277848E-2,-9.869006E-4],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false,false,false],"id":152,"left_children":[1,-1,3,-1,5,-1,-1],"loss_changes":[2.879397E-2,0E0,4.5362543E-2,0E0,2.0106424E-2,0E0,0E0],"parents":[2147483647,0,0,2,2,4,4],"right_children":[2,-1,4,-1,6,-1,-1],"split_conditions":[-7.922905E-1,2.9548613E-2,-4.3129416E0,2.3654656E-2,-7.277336E-1,-5.0277848E-2,-9.869006E-4],"split_indices":[443,0,136,0,92,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[4.587796E0,1.0628893E0,3.5249069E0,1.1629081E0,2.3619988E0,1.1640296E0,1.1979692E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[7.7536004E-4,-2.7838238E-2,4.5488223E-2,3.905701E-2,-1.9518847E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false],"id":153,"left_children":[1,-1,3,-1,-1],"loss_changes":[2.7492417E-2,0E0,4.999759E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-5.281218E0,-2.7838238E-2,5.081134E-1,3.905701E-2,-1.9518847E-2],"split_indices":[13,0,29,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.562312E0,1.127496E0,3.4348161E0,1.8441908E0,1.5906253E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-9.2456886E-4,2.0952854E-2,-6.049767E-2,-4.5082375E-2,1.4505329E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":154,"left_children":[1,-1,3,-1,-1],"loss_changes":[2.7588198E-2,0E0,4.190484E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-1.2516327E0,2.0952854E-2,-2.057909E-1,-4.5082375E-2,1.4505329E-2],"split_indices":[26,0,319,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.541695E0,1.9970236E0,2.5446715E0,1.185942E0,1.3587295E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-2.438307E-3,-4.663286E-2,2.736809E-2,-3.6772124E-2,1.8316045E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false],"id":155,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.7116697E-2,4.3582074E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[5.5464067E0,1.551867E0,2.736809E-2,-3.6772124E-2,1.8316045E-2],"split_indices":[70,3,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.523023E0,3.4142115E0,1.1088117E0,1.921156E0,1.4930555E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-1.1792238E-4,-4.3454394E-2,2.8202705E-2,-1.014024E-1,2.2671398E-2,-5.0850473E-2,8.568981E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,true,false,false,false],"id":156,"left_children":[1,3,-1,5,-1,-1,-1],"loss_changes":[2.6536616E-2,3.837473E-2,0E0,3.512577E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,3,3],"right_children":[2,4,-1,6,-1,-1,-1],"split_conditions":[5.923786E0,3.7927504E0,2.8202705E-2,2.1395693E0,2.2671398E-2,-5.0850473E-2,8.568981E-3],"split_indices":[161,73,0,113,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[4.503996E0,3.4526997E0,1.0512964E0,2.4217343E0,1.0309652E0,1.3894763E0,1.032258E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[1.4173669E-3,5.982019E-2,-2.03803E-2,-1.2473094E-2,3.7477266E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":157,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.6246974E-2,2.8330952E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[1.3957454E-1,-7.263844E-1,-2.03803E-2,-1.2473094E-2,3.7477266E-2],"split_indices":[226,114,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.476863E0,2.5048783E0,1.9719846E0,1.120741E0,1.3841373E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.8684119E-3,-5.717941E-2,2.0838639E-2,2.0203492E-2,-4.2089E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":158,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.577298E-2,4.5300134E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[1.5412903E0,-2.8239298E0,2.0838639E-2,2.0203492E-2,-4.2089E-2],"split_indices":[137,63,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.462227E0,2.4639058E0,1.998321E0,1.0622407E0,1.4016651E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.0816015E-3,-6.1995085E-2,6.490103E-2,-4.0751155E-2,1.09494645E-2,-1.039141E-2,3.8779475E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false,false,false],"id":159,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[2.5934791E-2,2.8587077E-2,2.4941595E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[-3.731128E-2,-1.8466124E-1,-4.677999E-2,-4.0751155E-2,1.09494645E-2,-1.039141E-2,3.8779475E-2],"split_indices":[434,362,346,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[4.443087E0,2.2489233E0,2.1941638E0,1.068614E0,1.1803093E0,1.0429997E0,1.151164E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[4.948474E-4,5.7581216E-2,-2.0786822E-2,-1.6346192E-2,3.984364E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":160,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.558787E-2,3.6983237E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[2.1315716E-1,-1.4910538E0,-2.0786822E-2,-1.6346192E-2,3.984364E-2],"split_indices":[280,17,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.4215775E0,2.528241E0,1.8933367E0,1.1262472E0,1.4019938E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[8.825424E-4,-5.6946408E-2,2.0982016E-2,-4.498539E-2,2.0152923E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false],"id":161,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.5566898E-2,5.0393336E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[5.617125E-1,8.4881775E-2,2.0982016E-2,-4.498539E-2,2.0152923E-2],"split_indices":[158,298,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.404767E0,2.4927275E0,1.9120395E0,1.3060328E0,1.1866947E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-1.4287245E-4,-4.323173E-2,2.6989948E-2,-3.8715772E-2,1.3102499E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":162,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.4795448E-2,3.7841827E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.9678495E0,-1.7314615E0,2.6989948E-2,-3.8715772E-2,1.3102499E-2],"split_indices":[6,76,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.384395E0,3.3180075E0,1.0663877E0,1.4254315E0,1.892576E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.2936135E-4,-2.4423258E-2,4.892658E-2,4.2890474E-2,-1.177053E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,false],"id":163,"left_children":[1,-1,3,-1,-1],"loss_changes":[2.5340278E-2,0E0,3.862859E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.208375E0,-2.4423258E-2,-2.6147842E0,4.2890474E-2,-1.177053E-2],"split_indices":[153,0,75,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.367547E0,1.3849475E0,2.9825995E0,1.1423744E0,1.8402251E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.3799561E-3,5.6173015E-2,-2.1001298E-2,-1.6595727E-2,3.6504976E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":164,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.4859194E-2,3.2891694E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[2.1612035E-1,-2.0657914E0,-2.1001298E-2,-1.6595727E-2,3.6504976E-2],"split_indices":[258,94,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.350269E0,2.5817003E0,1.7685686E0,1.0130879E0,1.5686125E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[7.903033E-4,5.2330963E-2,-2.2378456E-2,4.2316563E-2,-1.13214925E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":165,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.463109E-2,3.5339817E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.0724077E0,-2.2209193E-1,-2.2378456E-2,4.2316563E-2,-1.13214925E-2],"split_indices":[25,274,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.3347063E0,2.7561507E0,1.5785556E0,1.1032757E0,1.6528751E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[2.2122073E-3,5.4728646E-2,-2.344376E-2,-1.4919007E-2,3.7931874E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":166,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.672108E-2,3.4786742E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[2.2912774E0,-1.3288934E0,-2.344376E-2,-1.4919007E-2,3.7931874E-2],"split_indices":[81,161,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.318395E0,2.8045058E0,1.5138892E0,1.2663629E0,1.5381429E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.5005202E-3,-2.5728479E-2,4.7318418E-2,-1.7977035E-2,3.586682E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":167,"left_children":[1,-1,3,-1,-1],"loss_changes":[2.5255648E-2,0E0,3.9088413E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-5.318883E0,-2.5728479E-2,-1.4493701E0,-1.7977035E-2,3.586682E-2],"split_indices":[160,0,5,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.3018174E0,1.1809175E0,3.1209E0,1.3247231E0,1.7961768E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-4.072562E-4,-5.738415E-2,2.1707203E-2,-3.9784804E-2,1.7943451E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":168,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.6061242E-2,3.8980383E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[1.7952664E0,1.3743892E0,2.1707203E-2,-3.9784804E-2,1.7943451E-2],"split_indices":[133,66,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.2845173E0,2.5214882E0,1.763029E0,1.4555458E0,1.0659424E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-3.4456825E-4,2.160093E-2,-5.6428514E-2,-3.7508447E-2,9.659254E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,false],"id":169,"left_children":[1,-1,3,-1,-1],"loss_changes":[2.543425E-2,0E0,2.5472648E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-1.248637E0,2.160093E-2,1.799829E-1,-3.7508447E-2,9.659254E-3],"split_indices":[86,0,32,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.267036E0,1.7393931E0,2.527643E0,1.1932712E0,1.3343718E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-1.4570883E-3,-4.750311E-2,2.5297815E-2,-3.276675E-2,1.2748413E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":170,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.4756E-2,2.7182188E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[5.0954676E-1,1.5412903E0,2.5297815E-2,-3.276675E-2,1.2748413E-2],"split_indices":[487,137,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.2532654E0,3.058037E0,1.1952282E0,1.6872971E0,1.3707399E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-3.046753E-5,6.256531E-2,-6.439321E-2,-1.1762925E-2,4.206778E-2,-3.946868E-2,9.686416E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false,false,false],"id":171,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[2.5137084E-2,3.0500762E-2,2.4223367E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[-1.420579E-1,4.127674E-1,-5.5202925E-1,-1.1762925E-2,4.206778E-2,-3.946868E-2,9.686416E-3],"split_indices":[147,16,51,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[4.239305E0,2.1633105E0,2.0759945E0,1.150579E0,1.0127316E0,1.0120747E0,1.0639198E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[8.5831336E-5,-4.6085183E-2,2.5767209E-2,1.6428033E-2,-4.1411422E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":172,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.4628682E-2,4.5050077E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.8640652E0,3.6828887E-1,2.5767209E-2,1.6428033E-2,-4.1411422E-2],"split_indices":[156,136,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.2175603E0,3.0430324E0,1.1745278E0,1.6442499E0,1.3987825E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.7530455E-4,4.4010676E-2,-2.6766527E-2,-1.4228038E-2,3.053466E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":173,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.4299368E-2,2.6791569E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[6.333127E0,-1.3897004E0,-2.6766527E-2,-1.4228038E-2,3.053466E-2],"split_indices":[110,36,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.1987596E0,3.1579633E0,1.0407963E0,1.2920473E0,1.865916E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-1.884166E-3,4.410948E-2,-2.5836263E-2,3.2434326E-2,-1.5160394E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":174,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.3887219E-2,2.9785722E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[5.6943715E-1,2.0486091E-1,-2.5836263E-2,3.2434326E-2,-1.5160394E-2],"split_indices":[425,497,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.184987E0,3.015094E0,1.1698929E0,1.7137867E0,1.3013073E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-2.9649339E-3,-2.153227E-2,5.299723E-2,-1.9599218E-2,4.011316E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":175,"left_children":[1,-1,3,-1,-1],"loss_changes":[2.371444E-2,0E0,4.146309E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-1.7813606E0,-2.153227E-2,-2.0247204E0,-1.9599218E-2,4.011316E-2],"split_indices":[10,0,55,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.170623E0,1.7438682E0,2.4267545E0,1.0613574E0,1.3653971E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-2.247104E-3,2.400392E-2,-4.912683E-2,1.1478214E-2,-4.1664578E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false],"id":176,"left_children":[1,-1,3,-1,-1],"loss_changes":[2.380268E-2,0E0,3.591898E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.6159916E0,2.400392E-2,2.246585E-1,1.1478214E-2,-4.1664578E-2],"split_indices":[159,0,226,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.1530523E0,1.2510512E0,2.9020011E0,1.761088E0,1.1409131E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-2.8178233E-3,2.3261148E-2,-5.049397E-2,2.0954255E-2,-3.7162326E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":177,"left_children":[1,-1,3,-1,-1],"loss_changes":[2.3594258E-2,0E0,4.247009E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.5090895E0,2.3261148E-2,-2.4530811E0,2.0954255E-2,-3.7162326E-2],"split_indices":[164,0,61,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.1367836E0,1.3072205E0,2.8295631E0,1.0900549E0,1.7395083E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-6.293529E-4,-2.4549548E-2,4.937732E-2,-1.5564592E-2,3.708734E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":178,"left_children":[1,-1,3,-1,-1],"loss_changes":[2.4826817E-2,0E0,3.4920365E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.5643492E0,-2.4549548E-2,-3.4882784E-1,-1.5564592E-2,3.708734E-2],"split_indices":[154,0,11,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.118911E0,1.3272542E0,2.7916565E0,1.308429E0,1.4832275E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-1.010025E-3,4.4930603E-2,-2.6365303E-2,-1.0399259E-2,3.6270875E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":179,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.431405E-2,2.837545E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[5.099218E0,1.2116785E0,-2.6365303E-2,-1.0399259E-2,3.6270875E-2],"split_indices":[23,47,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.102743E0,2.9994118E0,1.1033312E0,1.7303168E0,1.2690951E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-7.85988E-4,-4.970158E-2,2.4292452E-2,1.6534086E-2,-3.316661E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":180,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.4375362E-2,3.0352669E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[2.0623732E0,-1.5987604E0,2.4292452E-2,1.6534086E-2,-3.316661E-2],"split_indices":[112,95,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.0884852E0,2.8033915E0,1.2850938E0,1.0643945E0,1.738997E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[9.406458E-4,4.9892318E-2,-2.2712134E-2,-1.3068268E-2,3.705917E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":181,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.281443E-2,3.0837344E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.4428743E-1,1.0090637E-1,-2.2712134E-2,-1.3068268E-2,3.705917E-2],"split_indices":[410,47,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.073839E0,2.6991053E0,1.374734E0,1.3695178E0,1.3295875E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[9.3922333E-4,6.1346672E-2,-5.985272E-2,2.90781E-2,-1.5651516E-3,-3.465252E-2,7.4538873E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,true,false,false,false,false],"id":182,"left_children":[1,3,5,-1,-1,-1,-1],"loss_changes":[2.2251276E-2,7.6855067E-3,1.7143091E-2,0E0,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1,2,2],"right_children":[2,4,6,-1,-1,-1,-1],"split_conditions":[7.6690674E-1,-2.4012692E-2,-4.964943E-1,2.90781E-2,-1.5651516E-3,-3.465252E-2,7.4538873E-3],"split_indices":[64,430,77,0,0,0,0],"split_type":[0,0,0,0,0,0,0],"sum_hessian":[4.0593853E0,2.0315547E0,2.0278306E0,1.0266366E0,1.0049181E0,1.0042119E0,1.0236187E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"7","size_leaf_vector":"0"}},{"base_weights":[7.110974E-4,-4.3337695E-2,2.5040777E-2,1.4965556E-2,-3.5453215E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":183,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.2013884E-2,3.3465724E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[6.299681E-1,3.56751E-2,2.5040777E-2,1.4965556E-2,-3.5453215E-2],"split_indices":[460,491,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.046506E0,2.9517453E0,1.0947607E0,1.4629207E0,1.4888246E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.3828814E-3,-4.4604324E-2,2.3952197E-2,1.3724844E-2,-3.4428716E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":184,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.171545E-2,2.9426124E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[2.9895558E0,-3.954602E-2,2.3952197E-2,1.3724844E-2,-3.4428716E-2],"split_indices":[155,258,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.03158E0,2.8137915E0,1.2177885E0,1.381946E0,1.4318455E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.5985863E-3,-5.8128897E-2,1.8389745E-2,-3.315634E-2,7.0130136E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":185,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.1455515E-2,1.5486649E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[-5.783028E-3,1.16206445E-1,1.8389745E-2,-3.315634E-2,7.0130136E-3],"split_indices":[264,488,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.0177917E0,2.0216029E0,1.9961889E0,1.013881E0,1.0077219E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-4.4389686E-4,-1.792719E-2,5.7762392E-2,-4.459141E-3,3.0443009E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,false],"id":186,"left_children":[1,-1,3,-1,-1],"loss_changes":[2.0740625E-2,0E0,1.1146912E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-2.0933533E-1,-1.792719E-2,-6.931937E-2,-4.459141E-3,3.0443009E-2],"split_indices":[56,0,438,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[4.007773E0,1.9718106E0,2.0359623E0,1.0129849E0,1.0229775E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.3128922E-3,2.5003701E-2,-4.04852E-2,1.8397793E-2,-2.9251441E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":187,"left_children":[1,-1,3,-1,-1],"loss_changes":[2.0515306E-2,0E0,2.9112075E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-5.9482584E0,2.5003701E-2,-4.8569068E-1,1.8397793E-2,-2.9251441E-2],"split_indices":[63,0,202,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.9991174E0,1.0143504E0,2.984767E0,1.0444063E0,1.9403607E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[3.7504847E-3,5.1367197E-2,-2.0136347E-2,3.7108753E-2,-1.480308E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":188,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.027321E-2,3.195282E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.810327E-1,-1.420579E-1,-2.0136347E-2,3.7108753E-2,-1.480308E-2],"split_indices":[238,147,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.9857807E0,2.5486326E0,1.4371481E0,1.3505E0,1.1981326E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[3.0982576E-3,-4.1196775E-2,2.4118103E-2,-3.25221E-2,1.9917157E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":189,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.0333828E-2,3.483151E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.0350685E0,2.0390704E0,2.4118103E-2,-3.25221E-2,1.9917157E-2],"split_indices":[6,73,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.9716268E0,2.8216622E0,1.1499645E0,1.7320343E0,1.0896279E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[4.4832383E-3,-3.8348086E-2,2.533299E-2,-3.1538095E-2,1.610741E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":190,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.0214079E-2,2.9709991E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.4581182E0,3.410101E-3,2.533299E-2,-3.1538095E-2,1.610741E-2],"split_indices":[2,31,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.956746E0,2.9154623E0,1.041284E0,1.607182E0,1.3082802E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[5.137703E-3,-2.2452652E-2,4.662738E-2,3.2393653E-2,-1.7485738E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":191,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.9893035E-2,0E0,3.1209249E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-5.318883E0,-2.2452652E-2,3.0724077E0,3.2393653E-2,-1.7485738E-2],"split_indices":[160,0,25,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.9432378E0,1.0722951E0,2.8709428E0,1.7931298E0,1.077813E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[2.0947775E-3,-2.3656907E-2,4.350702E-2,3.45334E-2,-1.2914432E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":192,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.9956078E-2,0E0,2.9161E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-4.649165E0,-2.3656907E-2,-2.1343412E-2,3.45334E-2,-1.2914432E-2],"split_indices":[90,0,184,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.9294343E0,1.0238568E0,2.9055777E0,1.4095677E0,1.49601E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.0769205E-3,4.5130678E-2,-2.2385431E-2,3.6282454E-2,-1.9765684E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":193,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.9761153E-2,3.9366636E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[4.738214E-1,2.0457262E-1,-2.2385431E-2,3.6282454E-2,-1.9765684E-2],"split_indices":[314,168,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.9161339E0,2.7306848E0,1.1854491E0,1.5695004E0,1.1611843E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-4.1582636E-4,-4.2141166E-2,2.4275025E-2,1.8054154E-2,-3.3586197E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":194,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.003956E-2,3.4569923E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[5.077216E0,-2.106997E0,2.4275025E-2,1.8054154E-2,-3.3586197E-2],"split_indices":[4,22,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.9002514E0,2.8962696E0,1.0039817E0,1.2306019E0,1.6656677E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.4788884E-3,5.1504277E-2,-2.210075E-2,3.6440805E-2,-1.0362002E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":195,"left_children":[1,3,-1,-1,-1],"loss_changes":[2.2160465E-2,2.5395716E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[2.2912774E0,-1.6307354E-1,-2.210075E-2,3.6440805E-2,-1.0362002E-2],"split_indices":[81,272,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.8855214E0,2.5215664E0,1.363955E0,1.1636565E0,1.3579099E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.878935E-3,-4.6002675E-2,2.1939315E-2,1.515612E-2,-3.1489007E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":196,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.9987844E-2,2.53813E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[4.6738672E-1,-1.8640366E0,2.1939315E-2,1.515612E-2,-3.1489007E-2],"split_indices":[177,117,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.8725817E0,2.5280757E0,1.344506E0,1.0129503E0,1.5151254E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[2.456951E-3,2.1235384E-2,-4.742802E-2,-3.5117608E-2,1.0496981E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,false],"id":197,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.9923752E-2,0E0,2.3654327E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-2.826052E-1,2.1235384E-2,-3.0751958E-2,-3.5117608E-2,1.0496981E-2],"split_indices":[363,0,357,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.8603392E0,1.4522412E0,2.408098E0,1.0774828E0,1.3306152E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.0052655E-3,-1.8258903E-2,5.636319E-2,3.8183693E-2,-1.3143488E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,false],"id":198,"left_children":[1,-1,3,-1,-1],"loss_changes":[2.0036202E-2,0E0,2.7443044E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-1.5458298E-1,-1.8258903E-2,8.289962E-1,3.8183693E-2,-1.3143488E-2],"split_indices":[84,0,64,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.8485432E0,1.7704427E0,2.0781004E0,1.0583235E0,1.0197769E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[2.1381297E-3,1.7941995E-2,-1.7398106E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":199,"left_children":[1,-1,-1],"loss_changes":[2.0236913E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[5.8762825E-1,1.7941995E-2,-1.7398106E-2],"split_indices":[132,0,0],"split_type":[0,0,0],"sum_hessian":[3.8356092E0,1.9606677E0,1.8749415E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.6116559E-3,-4.1058347E-2,2.3535827E-2,-3.0258328E-2,1.0123118E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":200,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.9058304E-2,2.0232486E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[5.0954676E-1,-1.629901E-1,2.3535827E-2,-3.0258328E-2,1.0123118E-2],"split_indices":[487,400,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.8301568E0,2.7620547E0,1.068102E0,1.3413221E0,1.4207326E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.8244641E-3,5.126393E-2,-1.967687E-2,3.5917155E-2,-1.3263889E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":201,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.9425318E-2,2.7135909E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.325308E-1,5.652153E-2,-1.967687E-2,3.5917155E-2,-1.3263889E-2],"split_indices":[275,489,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.819489E0,2.3417118E0,1.4777771E0,1.2159101E0,1.1258017E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.924685E-3,5.566814E-2,-1.8162427E-2,-1.1160577E-2,3.6456365E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":202,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.950782E-2,2.3509344E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[7.9127884E-1,-6.0569882E-2,-1.8162427E-2,-1.1160577E-2,3.6456365E-2],"split_indices":[37,56,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.8069396E0,2.104841E0,1.7020985E0,1.0537934E0,1.0510476E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[2.8641175E-3,2.2563782E-2,-4.412818E-2,1.2574571E-2,-3.725998E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":203,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.9621886E-2,0E0,2.9576346E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.1927748E0,2.2563782E-2,5.0466776E-1,1.2574571E-2,-3.725998E-2],"split_indices":[106,0,136,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.7952185E0,1.2579584E0,2.53726E0,1.4527233E0,1.0845368E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.5414656E-3,-5.1282328E-2,1.898593E-2,-3.8879476E-2,1.7034229E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":204,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.8841866E-2,3.3843584E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[5.617125E-1,8.4881775E-2,1.898593E-2,-3.8879476E-2,1.7034229E-2],"split_indices":[158,298,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.781817E0,2.1294703E0,1.6523465E0,1.1191281E0,1.0103422E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[4.5146077E-4,2.0084035E-2,-4.968893E-2,1.1686298E-2,-3.2819197E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":205,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.9222086E-2,0E0,2.168537E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-2.5694861E0,2.0084035E-2,-8.187554E-1,1.1686298E-2,-3.2819197E-2],"split_indices":[61,0,63,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.7675729E0,1.4755446E0,2.2920284E0,1.0623906E0,1.2296379E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[7.807054E-5,-4.6790145E-2,2.0252397E-2,1.4643494E-2,-3.2964755E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":206,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.8190544E-2,2.5851415E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.2208195E0,-1.5708787E0,2.0252397E-2,1.4643494E-2,-3.2964755E-2],"split_indices":[88,37,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.7564282E0,2.3966746E0,1.3597537E0,1.0428399E0,1.3538347E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[4.4257345E-4,1.9162001E-2,-4.9804337E-2,-3.071618E-2,7.005722E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":207,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.8302893E-2,0E0,1.4736373E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-1.248637E0,1.9162001E-2,1.799829E-1,-3.071618E-2,7.005722E-3],"split_indices":[86,0,32,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.7445168E0,1.5352528E0,2.209264E0,1.0529058E0,1.1563582E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-6.6397275E-4,1.6226543E-2,-1.7875452E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":208,"left_children":[1,-1,-1],"loss_changes":[1.8499872E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[5.6807205E-2,1.6226543E-2,-1.7875452E-2],"split_indices":[421,0,0],"split_type":[0,0,0],"sum_hessian":[3.7355013E0,1.9787527E0,1.7567486E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-8.1467297E-4,2.156985E-2,-4.3345857E-2,-3.5171065E-2,1.8046778E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,false],"id":209,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.774648E-2,0E0,3.465096E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-5.1557803E0,2.156985E-2,1.2712517E0,-3.5171065E-2,1.8046778E-2],"split_indices":[126,0,56,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.7305956E0,1.1219398E0,2.608656E0,1.4446213E0,1.1640346E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[3.4718646E-4,-5.094477E-2,1.8441604E-2,1.1651128E-2,-3.3944137E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":210,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.7918723E-2,2.176987E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[2.3680219E-1,5.7764053E-2,1.8441604E-2,1.1651128E-2,-3.3944137E-2],"split_indices":[247,121,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.716442E0,2.1113122E0,1.60513E0,1.0178283E0,1.0934838E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-5.9573737E-4,-1.9211264E-2,4.9239185E-2,-8.535443E-3,3.1355187E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false],"id":211,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.80307E-2,0E0,1.6719723E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-2.0963696E-1,-1.9211264E-2,6.0760444E-1,-8.535443E-3,3.1355187E-2],"split_indices":[454,0,11,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.7056746E0,1.5048932E0,2.2007813E0,1.1166656E0,1.0841157E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[3.655137E-4,-1.6110769E-2,1.7934965E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":212,"left_children":[1,-1,-1],"loss_changes":[1.8297557E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.7847767E-2,-1.6110769E-2,1.7934965E-2],"split_indices":[19,0,0],"split_type":[0,0,0],"sum_hessian":[3.69621E0,1.9855831E0,1.7106268E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[6.457561E-4,4.4705667E-2,-2.0607116E-2,-1.0404788E-2,3.540957E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":213,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.7402545E-2,2.4318194E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[4.1914754E0,5.617125E-1,-2.0607116E-2,-1.0404788E-2,3.540957E-2],"split_indices":[110,158,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.6913123E0,2.4742644E0,1.2170479E0,1.4410684E0,1.033196E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.415801E-3,4.602085E-2,-2.0576509E-2,-1.266127E-2,3.2493666E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":214,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.7769745E-2,2.3569662E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.4605944E-1,-1.8168974E-1,-2.0576509E-2,-1.266127E-2,3.2493666E-2],"split_indices":[368,11,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.679925E0,2.4569874E0,1.2229376E0,1.1502811E0,1.3067063E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[8.525792E-4,-1.6022014E-2,1.8161148E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":215,"left_children":[1,-1,-1],"loss_changes":[1.835239E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.1221077E-1,-1.6022014E-2,1.8161148E-2],"split_indices":[76,0,0],"split_type":[0,0,0],"sum_hessian":[3.6686945E0,1.9767835E0,1.691911E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.0576071E-3,1.6737068E-2,-1.6726157E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":216,"left_children":[1,-1,-1],"loss_changes":[1.7612195E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.3998775E-1,1.6737068E-2,-1.6726157E-2],"split_indices":[159,0,0],"split_type":[0,0,0],"sum_hessian":[3.663758E0,1.8751751E0,1.7885829E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[7.4319175E-4,2.1103118E-2,-4.4823814E-2,-3.4074645E-2,1.5031386E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":217,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.7928975E-2,0E0,2.8055582E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.5600772E0,2.1103118E-2,2.4339342E-1,-3.4074645E-2,1.5031386E-2],"split_indices":[85,0,73,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.6589804E0,1.2325702E0,2.4264102E0,1.2932115E0,1.1331987E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[9.439596E-4,1.6511638E-2,-1.7047726E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":218,"left_children":[1,-1,-1],"loss_changes":[1.7648153E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[4.0255515E-3,1.6511638E-2,-1.7047726E-2],"split_indices":[260,0,0],"split_type":[0,0,0],"sum_hessian":[3.6465073E0,1.9075624E0,1.7389449E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[5.462319E-4,-1.6043002E-2,1.7848685E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":219,"left_children":[1,-1,-1],"loss_changes":[1.7963935E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.0041963E0,-1.6043002E-2,1.7848685E-2],"split_indices":[78,0,0],"split_type":[0,0,0],"sum_hessian":[3.6417851E0,1.9487406E0,1.6930445E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[7.7308307E-4,-4.0542662E-2,2.234511E-2,1.3216374E-2,-3.203362E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":220,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.7141143E-2,2.4982506E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.8640652E0,7.6690674E-1,2.234511E-2,1.3216374E-2,-3.203362E-2],"split_indices":[156,64,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.6369386E0,2.6189556E0,1.0179828E0,1.2971326E0,1.321823E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.2888993E-3,1.9689558E-2,-4.5079984E-2,1.4165773E-2,-3.2833513E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":221,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.675876E-2,0E0,2.465989E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-2.64045E-1,1.9689558E-2,-1.587546E-1,1.4165773E-2,-3.2833513E-2],"split_indices":[390,0,272,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.6254478E0,1.3444358E0,2.2810118E0,1.04659E0,1.2344218E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.1073869E-3,-5.0528508E-2,1.990507E-2,1.0516313E-2,-3.387952E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false],"id":222,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.8897451E-2,2.074661E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[1.7952664E0,-1.2247467E-1,1.990507E-2,1.0516313E-2,-3.387952E-2],"split_indices":[133,34,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.614266E0,2.143412E0,1.4708537E0,1.0886472E0,1.0547649E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[5.8847386E-6,-4.291584E-2,2.095776E-2,1.281584E-2,-3.2778572E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":223,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.6800353E-2,2.4387604E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[2.0623732E0,-6.119106E-1,2.095776E-2,1.281584E-2,-3.2778572E-2],"split_indices":[112,29,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.603486E0,2.4708738E0,1.1326121E0,1.2340943E0,1.2367796E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[2.0255559E-4,-4.32086E-2,2.1151613E-2,-3.3424E-2,1.4251347E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false],"id":224,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.7061936E-2,2.669496E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[2.674161E0,4.1862488E-2,2.1151613E-2,-3.3424E-2,1.4251347E-2],"split_indices":[155,77,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.5923443E0,2.4592066E0,1.1331378E0,1.2735008E0,1.1857058E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[5.5472367E-4,-2.0290833E-2,4.4979744E-2,3.1718384E-2,-1.1251771E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":225,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.691839E-2,0E0,2.0872928E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-4.0247116E0,-2.0290833E-2,7.3667336E-1,3.1718384E-2,-1.1251771E-2],"split_indices":[94,0,64,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.5805216E0,1.2063456E0,2.374176E0,1.2049745E0,1.1692015E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[4.4362803E-4,-1.6220182E-2,1.657251E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":226,"left_children":[1,-1,-1],"loss_changes":[1.6638411E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.731128E-2,-1.6220182E-2,1.657251E-2],"split_indices":[434,0,0],"split_type":[0,0,0],"sum_hessian":[3.5702124E0,1.7964816E0,1.7737308E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[3.8904086E-4,-4.5179896E-2,1.9425346E-2,-3.23869E-2,1.2484647E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":227,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.6316468E-2,2.2267263E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.4950408E-1,-3.339064E-1,1.9425346E-2,-3.23869E-2,1.2484647E-2],"split_indices":[332,19,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.5657735E0,2.2621717E0,1.3036019E0,1.1712444E0,1.0909274E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.4981205E-4,2.0943688E-2,-4.2440478E-2,-3.213224E-2,1.4266144E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":228,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.6478281E-2,0E0,2.5164755E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-4.4137E-1,2.0943688E-2,1.4613691E-1,-3.213224E-2,1.4266144E-2],"split_indices":[295,0,228,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.5553303E0,1.1064365E0,2.4488938E0,1.3143127E0,1.1345811E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[7.6535944E-4,4.6921954E-2,-1.8944496E-2,-1.1826397E-2,3.1780194E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":229,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.6368186E-2,2.0543523E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.2464916E-1,-5.1303856E-2,-1.8944496E-2,-1.1826397E-2,3.1780194E-2],"split_indices":[291,346,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.5440583E0,2.2122734E0,1.331785E0,1.032927E0,1.1793463E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[7.7631284E-4,-1.6886713E-2,1.5248271E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":230,"left_children":[1,-1,-1],"loss_changes":[1.5811156E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-9.9897707E-1,-1.6886713E-2,1.5248271E-2],"split_indices":[161,0,0],"split_type":[0,0,0],"sum_hessian":[3.5340505E0,1.5930849E0,1.9409655E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.8129151E-4,-2.0309636E-2,4.2081848E-2,-1.2883744E-2,3.0527277E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false],"id":231,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.5732575E-2,0E0,2.1711033E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.208375E0,-2.0309636E-2,-4.559751E-2,-1.2883744E-2,3.0527277E-2],"split_indices":[153,0,182,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.529789E0,1.1122352E0,2.417554E0,1.1126155E0,1.3049384E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-4.4796895E-4,-1.8798804E-2,4.5397926E-2,3.2256342E-2,-1.0603825E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":232,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.5736038E-2,0E0,1.976166E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.3287156E-1,-1.8798804E-2,-1.016685E0,3.2256342E-2,-1.0603825E-2],"split_indices":[366,0,91,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.5196266E0,1.3376167E0,2.18201E0,1.0457759E0,1.136234E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[3.3613315E-4,-4.463378E-2,1.9072257E-2,1.2211848E-2,-3.4932774E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":233,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.566317E-2,2.4492351E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.081163E0,1.1673189E-1,1.9072257E-2,1.2211848E-2,-3.4932774E-2],"split_indices":[100,280,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.5100107E0,2.2232237E0,1.2867872E0,1.2138159E0,1.0094078E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-5.977492E-4,4.002297E-2,-2.0018125E-2,3.0025797E-2,-1.3986589E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":234,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.4756108E-2,2.2423582E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[4.9670988E-1,1.9353487E-1,-2.0018125E-2,3.0025797E-2,-1.3986589E-2],"split_indices":[351,314,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.4990692E0,2.412153E0,1.0869161E0,1.332987E0,1.079166E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-1.2919676E-3,1.47589715E-2,-1.6174948E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":235,"left_children":[1,-1,-1],"loss_changes":[1.4579512E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[9.095875E-2,1.47589715E-2,-1.6174948E-2],"split_indices":[189,0,0],"split_type":[0,0,0],"sum_hessian":[3.4890091E0,1.8138864E0,1.6751227E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.2219293E-3,1.9102154E-2,-4.2533074E-2,-3.0357825E-2,1.27636595E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":236,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.47324605E-2,0E0,2.0974532E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.1927748E0,1.9102154E-2,2.1221077E-1,-3.0357825E-2,1.27636595E-2],"split_indices":[106,0,76,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.4851103E0,1.1450402E0,2.3400702E0,1.2729788E0,1.0670915E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-5.3566205E-4,4.17299E-2,-1.8488433E-2,-1.453071E-2,3.297633E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":237,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.4127583E-2,2.5254684E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[4.198103E-1,-3.477621E-2,-1.8488433E-2,-1.453071E-2,3.297633E-2],"split_indices":[452,100,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.4753563E0,2.2415347E0,1.2338216E0,1.0899969E0,1.1515378E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-3.659463E-4,-2.1377943E-2,4.4711854E-2,-1.1884059E-2,2.9633576E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false],"id":238,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.7454043E-2,0E0,1.9157642E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.5643492E0,-2.1377943E-2,-1.4890568E-1,-1.1884059E-2,2.9633576E-2],"split_indices":[154,0,174,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.4646573E0,1.1209387E0,2.3437185E0,1.0200754E0,1.3236431E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-1.0584736E-3,-4.828721E-2,1.796671E-2,-2.8163627E-2,6.126653E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":239,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.5714608E-2,1.151365E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[1.0823454E-1,-1.2594619E0,1.796671E-2,-2.8163627E-2,6.126653E-3],"split_indices":[487,112,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.4546986E0,2.0634434E0,1.3912553E0,1.020189E0,1.0432544E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-1.6686609E-3,1.5102593E-2,-1.652843E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":240,"left_children":[1,-1,-1],"loss_changes":[1.5131345E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[7.7981496E-1,1.5102593E-2,-1.652843E-2],"split_indices":[69,0,0],"split_type":[0,0,0],"sum_hessian":[3.447278E0,1.7760295E0,1.6712486E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.490507E-3,1.6378049E-2,-1.4900416E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":241,"left_children":[1,-1,-1],"loss_changes":[1.4717009E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.9723969E-1,1.6378049E-2,-1.4900416E-2],"split_indices":[95,0,0],"split_type":[0,0,0],"sum_hessian":[3.443192E0,1.5295053E0,1.9136868E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-7.1103213E-4,-1.6123284E-2,1.4842089E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":242,"left_children":[1,-1,-1],"loss_changes":[1.447379E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-5.077629E-1,-1.6123284E-2,1.4842089E-2],"split_indices":[47,0,0],"split_type":[0,0,0],"sum_hessian":[3.4392505E0,1.637681E0,1.8015695E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-7.878374E-4,1.9329205E-2,-4.1470338E-2,-3.0439671E-2,1.2637822E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,false],"id":243,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.44400755E-2,0E0,2.1025188E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.8978767E0,1.9329205E-2,2.4339342E-1,-3.0439671E-2,1.2637822E-2],"split_indices":[85,0,73,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.4353535E0,1.0954623E0,2.3398912E0,1.2377973E0,1.1020939E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-2.2290483E-4,-1.985186E-2,4.003972E-2,3.1722173E-2,-1.6442891E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false],"id":244,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.4401152E-2,0E0,2.6802856E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-5.245779E-1,-1.985186E-2,5.8195466E-1,3.1722173E-2,-1.6442891E-2],"split_indices":[235,0,235,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.4256878E0,1.0546485E0,2.3710394E0,1.3329183E0,1.0381211E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-1.1749158E-3,-4.2220093E-2,1.925459E-2,-3.0945431E-2,1.2865969E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":245,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.4552196E-2,2.1613352E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.7718365E0,6.062341E-2,1.925459E-2,-3.0945431E-2,1.2865969E-2],"split_indices":[36,382,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.4145584E0,2.3148205E0,1.0997379E0,1.2254472E0,1.0893734E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-5.737743E-4,-4.620201E-2,1.7276699E-2,-2.5458412E-2,4.5457925E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":246,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.4350182E-2,8.45663E-3,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[1.7952664E0,-6.199204E-1,1.7276699E-2,-2.5458412E-2,4.5457925E-3],"split_indices":[133,65,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.4046912E0,2.0231717E0,1.3815196E0,1.0061018E0,1.0170698E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-1.2361491E-3,1.756522E-2,-1.4518963E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":247,"left_children":[1,-1,-1],"loss_changes":[1.5235485E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.248637E0,1.756522E-2,-1.4518963E-2],"split_indices":[86,0,0],"split_type":[0,0,0],"sum_hessian":[3.398478E0,1.3921179E0,2.0063603E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.7196001E-4,-1.7205207E-2,1.3807586E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":248,"left_children":[1,-1,-1],"loss_changes":[1.4247231E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.7813606E0,-1.7205207E-2,1.3807586E-2],"split_indices":[10,0,0],"split_type":[0,0,0],"sum_hessian":[3.394341E0,1.4089905E0,1.9853505E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-8.959096E-4,1.8091539E-2,-4.2212646E-2,1.0203726E-2,-3.0722918E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":249,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.3647651E-2,0E0,1.8209148E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.947982E0,1.8091539E-2,3.7443593E-2,1.0203726E-2,-3.0722918E-2],"split_indices":[162,0,258,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.3905292E0,1.1812315E0,2.2092977E0,1.166804E0,1.0424937E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-1.1982742E-3,4.166961E-2,-1.8205002E-2,2.6200878E-2,-8.054481E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":250,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.370238E-2,1.2168301E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.0724077E0,4.548741E-2,-1.8205002E-2,2.6200878E-2,-8.054481E-3],"split_indices":[25,324,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.3818555E0,2.1395118E0,1.2423437E0,1.1190348E0,1.020477E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-8.4518903E-4,4.070796E-2,-1.855257E-2,2.6508559E-2,-8.855226E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":251,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.3605858E-2,1.3334384E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.4309313E-1,7.287896E-2,-1.855257E-2,2.6508559E-2,-8.855226E-3],"split_indices":[193,443,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.3748045E0,2.2051783E0,1.1696262E0,1.1598562E0,1.0453221E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-7.894725E-4,1.4148078E-2,-1.5856592E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":252,"left_children":[1,-1,-1],"loss_changes":[1.3396318E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.18710145E-1,1.4148078E-2,-1.5856592E-2],"split_indices":[269,0,0],"split_type":[0,0,0],"sum_hessian":[3.3674994E0,1.8020918E0,1.5654075E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-9.4550656E-4,-1.590313E-2,1.40603045E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":253,"left_children":[1,-1,-1],"loss_changes":[1.3347999E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.0247204E0,-1.590313E-2,1.40603045E-2],"split_indices":[55,0,0],"split_type":[0,0,0],"sum_hessian":[3.3639297E0,1.5583289E0,1.8056009E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.0862302E-3,4.2169645E-2,-1.7241206E-2,2.9543491E-2,-1.0466619E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":254,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.3058195E-2,1.664476E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.325308E-1,-4.7785673E-2,-1.7241206E-2,2.9543491E-2,-1.0466619E-2],"split_indices":[275,168,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.3603342E0,2.0442066E0,1.3161275E0,1.020517E0,1.0236896E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-3.2587306E-4,1.7207013E-2,-4.4435076E-2,-2.6061697E-2,6.1755828E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":255,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.3622182E-2,0E0,1.0148263E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-2.5694861E0,1.7207013E-2,-3.775251E-2,-2.6061697E-2,6.1755828E-3],"split_indices":[61,0,38,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.3522336E0,1.3224726E0,2.029761E0,1.0248119E0,1.0049492E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-8.5417257E-4,3.9702076E-2,-1.857879E-2,-1.1237571E-2,3.006622E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":256,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.3222833E-2,1.8794479E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.4428743E-1,1.0090637E-1,-1.857879E-2,-1.1237571E-2,3.006622E-2],"split_indices":[410,47,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.345663E0,2.220868E0,1.1247951E0,1.1437017E0,1.0771664E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-5.266807E-4,-1.693263E-2,1.3285345E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":257,"left_children":[1,-1,-1],"loss_changes":[1.3366278E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.5414672E-1,-1.693263E-2,1.3285345E-2],"split_indices":[218,0,0],"split_type":[0,0,0],"sum_hessian":[3.3369322E0,1.3690612E0,1.967871E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.2127403E-3,-4.0990774E-2,1.8564735E-2,-2.9643297E-2,9.806402E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":258,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.34124765E-2,1.7115023E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[4.5117092E-1,-7.105297E-1,1.8564735E-2,-2.9643297E-2,9.806402E-3],"split_indices":[214,161,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.333356E0,2.2593162E0,1.0740398E0,1.0747721E0,1.1845441E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-1.1586534E-3,4.2261887E-2,-1.7157627E-2,-7.985557E-4,1.9873064E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":259,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.29398275E-2,3.4211054E-3,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.0193767E-1,-2.2697258E-1,-1.7157627E-2,-7.985557E-4,1.9873064E-2],"split_indices":[448,82,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.32502E0,2.0118234E0,1.3131967E0,1.0095966E0,1.0022268E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-3.6674863E-4,1.4380137E-2,-1.490949E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":260,"left_children":[1,-1,-1],"loss_changes":[1.2677051E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.1725941E-2,1.4380137E-2,-1.490949E-2],"split_indices":[491,0,0],"split_type":[0,0,0],"sum_hessian":[3.320581E0,1.6921399E0,1.6284411E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-3.7980874E-4,-1.5244032E-2,1.4025104E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":261,"left_children":[1,-1,-1],"loss_changes":[1.2637283E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.462903E-1,-1.5244032E-2,1.4025104E-2],"split_indices":[400,0,0],"split_type":[0,0,0],"sum_hessian":[3.3171964E0,1.564686E0,1.7525104E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-5.4877234E-4,1.5714966E-2,-1.3828634E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":262,"left_children":[1,-1,-1],"loss_changes":[1.281462E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.7238207E-1,1.5714966E-2,-1.3828634E-2],"split_indices":[428,0,0],"split_type":[0,0,0],"sum_hessian":[3.3137817E0,1.4632124E0,1.8505694E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[5.3382795E-5,-1.8464636E-2,3.901015E-2,3.0933019E-2,-1.4112345E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":263,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.27450805E-2,0E0,2.2791218E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-4.582383E-1,-1.8464636E-2,4.486317E-2,3.0933019E-2,-1.4112345E-2],"split_indices":[446,0,265,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.3103433E0,1.0577701E0,2.252573E0,1.1773337E0,1.0752393E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-3.283658E-4,-3.9133392E-2,1.843235E-2,1.5357219E-2,-3.0904496E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":264,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.2712959E-2,2.4038916E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[2.674161E0,-8.5044146E-1,1.843235E-2,1.5357219E-2,-3.0904496E-2],"split_indices":[155,117,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.3006654E0,2.2522268E0,1.0484385E0,1.0153106E0,1.2369162E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[4.2524867E-4,4.148069E-2,-1.6796377E-2,2.8026175E-2,-9.176906E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":265,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.2259853E-2,1.4303957E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.622104E-1,7.5094607E-3,-1.6796377E-2,2.8026175E-2,-9.176906E-3],"split_indices":[349,226,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.29065E0,2.0577748E0,1.2328751E0,1.0237411E0,1.0340337E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[7.894914E-4,-3.7247766E-2,1.832684E-2,1.29602635E-2,-3.204694E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":266,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.2100073E-2,2.2793548E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[2.9614024E0,3.6828887E-1,1.832684E-2,1.29602635E-2,-3.204694E-2],"split_indices":[71,136,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.2834113E0,2.247794E0,1.0356174E0,1.2182422E0,1.0295517E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[3.575604E-4,-1.7687658E-2,3.9618157E-2,2.832052E-2,-1.0536219E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":267,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.2288758E-2,0E0,1.6294021E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-4.0247116E0,-1.7687658E-2,7.3667336E-1,2.832052E-2,-1.0536219E-2],"split_indices":[94,0,64,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.273843E0,1.1040583E0,2.1697848E0,1.1002299E0,1.0695549E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[2.8616362E-4,-1.5768062E-2,1.9551186E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":268,"left_children":[1,-1,-1],"loss_changes":[1.8053131E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.2456741E0,-1.5768062E-2,1.9551186E-2],"split_indices":[135,0,0],"split_type":[0,0,0],"sum_hessian":[3.2660096E0,1.9046626E0,1.361347E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.0904653E-3,-1.4302894E-2,1.8781854E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":269,"left_children":[1,-1,-1],"loss_changes":[1.5768139E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.86212E0,-1.4302894E-2,1.8781854E-2],"split_indices":[154,0,0],"split_type":[0,0,0],"sum_hessian":[3.2612112E0,1.9445983E0,1.3166128E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.8185073E-3,1.4725911E-2,-1.4043216E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":270,"left_children":[1,-1,-1],"loss_changes":[1.208276E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-4.3729258E-1,1.4725911E-2,-1.4043216E-2],"split_indices":[98,0,0],"split_type":[0,0,0],"sum_hessian":[3.2569363E0,1.6468189E0,1.6101174E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.434978E-3,-1.6609656E-2,4.2119287E-2,-1.0344373E-2,2.9004576E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,false],"id":271,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.2161749E-2,0E0,1.6066188E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.6237645E-1,-1.6609656E-2,6.324176E-2,-1.0344373E-2,2.9004576E-2],"split_indices":[495,0,409,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.253684E0,1.2072927E0,2.0463915E0,1.0043818E0,1.0420097E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.503415E-3,-1.3223415E-2,1.589446E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":272,"left_children":[1,-1,-1],"loss_changes":[1.2298369E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.5412903E0,-1.3223415E-2,1.589446E-2],"split_indices":[137,0,0],"split_type":[0,0,0],"sum_hessian":[3.2460124E0,1.7978511E0,1.4481614E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.6838558E-3,4.1001327E-2,-1.6521629E-2,-8.764915E-3,2.7959052E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false],"id":273,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.1725591E-2,1.4009846E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.810327E-1,6.1326504E-2,-1.6521629E-2,-8.764915E-3,2.7959052E-2],"split_indices":[238,76,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.2426887E0,2.0796232E0,1.1630656E0,1.0744438E0,1.0051794E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.8237968E-3,-1.4584406E-2,1.377186E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":274,"left_children":[1,-1,-1],"loss_changes":[1.16493255E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.9475713E-1,-1.4584406E-2,1.377186E-2],"split_indices":[84,0,0],"split_type":[0,0,0],"sum_hessian":[3.235586E0,1.461054E0,1.774532E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.0850123E-3,-3.6827493E-2,1.8196523E-2,-2.9140811E-2,1.3149343E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false],"id":275,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.1792546E-2,1.9865496E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[2.0623732E0,9.183703E-2,1.8196523E-2,-2.9140811E-2,1.3149343E-2],"split_indices":[112,211,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.2324471E0,2.2085965E0,1.0238507E0,1.1468271E0,1.0617694E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.2701348E-3,-1.4373428E-2,1.5272807E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":276,"left_children":[1,-1,-1],"loss_changes":[1.2750979E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-5.783028E-3,-1.4373428E-2,1.5272807E-2],"split_indices":[264,0,0],"split_type":[0,0,0],"sum_hessian":[3.2238538E0,1.6368762E0,1.5869776E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.103793E-3,-1.4592717E-2,1.3886695E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":277,"left_children":[1,-1,-1],"loss_changes":[1.1738191E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.0751958E-2,-1.4592717E-2,1.3886695E-2],"split_indices":[357,0,0],"split_type":[0,0,0],"sum_hessian":[3.220404E0,1.4964217E0,1.7239822E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[6.052545E-4,-1.625407E-2,1.2681983E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":278,"left_children":[1,-1,-1],"loss_changes":[1.19175175E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.5441092E-1,-1.625407E-2,1.2681983E-2],"split_indices":[268,0,0],"split_type":[0,0,0],"sum_hessian":[3.2172465E0,1.2601315E0,1.957115E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-4.2502105E-4,-1.3427057E-2,1.5366153E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":279,"left_children":[1,-1,-1],"loss_changes":[1.194073E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[5.617125E-1,-1.3427057E-2,1.5366153E-2],"split_indices":[158,0,0],"split_type":[0,0,0],"sum_hessian":[3.2140589E0,1.8012626E0,1.4127963E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.6336532E-4,4.0787723E-2,-1.6646164E-2,-5.4035924E-4,1.8886048E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":280,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.1782953E-2,2.9552318E-3,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.2464916E-1,-1.5451303E-2,-1.6646164E-2,-5.4035924E-4,1.8886048E-2],"split_indices":[291,402,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.2108479E0,2.0103714E0,1.2004764E0,1.0026456E0,1.0077258E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[5.20088E-4,1.6397953E-2,-1.2259828E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":281,"left_children":[1,-1,-1],"loss_changes":[1.1659699E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.3680513E-1,1.6397953E-2,-1.2259828E-2],"split_indices":[288,0,0],"split_type":[0,0,0],"sum_hessian":[3.2068222E0,1.2503878E0,1.9564344E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.3497412E-3,-1.6301846E-2,4.124209E-2,2.8050708E-2,-9.585205E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,false],"id":282,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.1579884E-2,0E0,1.4530522E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.435208E-1,-1.6301846E-2,3.3998775E-1,2.8050708E-2,-9.585205E-3],"split_indices":[366,0,159,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.2036905E0,1.1859566E0,2.0177338E0,1.0153117E0,1.0024221E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[1.4745627E-3,-3.8215075E-2,1.7015684E-2,7.6043666E-3,-2.4831034E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":283,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.1368968E-2,1.0706123E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.2208195E0,-7.591458E-3,1.7015684E-2,7.6043666E-3,-2.4831034E-2],"split_indices":[88,261,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.1965358E0,2.0395212E0,1.1570146E0,1.018129E0,1.0213922E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[9.76337E-4,1.6782036E-2,-1.255411E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":284,"left_children":[1,-1,-1],"loss_changes":[1.2204244E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.5694861E0,1.6782036E-2,-1.255411E-2],"split_indices":[61,0,0],"split_type":[0,0,0],"sum_hessian":[3.190497E0,1.2630605E0,1.9274365E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.6954424E-3,1.5503618E-2,-1.2543651E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":285,"left_children":[1,-1,-1],"loss_changes":[1.1266397E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.2528253E0,1.5503618E-2,-1.2543651E-2],"split_indices":[72,0,0],"split_type":[0,0,0],"sum_hessian":[3.187149E0,1.3957951E0,1.791354E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.8964971E-3,-1.5251026E-2,1.2942193E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":286,"left_children":[1,-1,-1],"loss_changes":[1.1293354E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.0963696E-1,-1.5251026E-2,1.2942193E-2],"split_indices":[454,0,0],"split_type":[0,0,0],"sum_hessian":[3.1841311E0,1.2953558E0,1.8887753E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[7.756212E-4,-3.8615998E-2,1.66768E-2,9.5388675E-3,-2.6785621E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":287,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.117442E-2,1.3623739E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.4950408E-1,-3.676063E-2,1.66768E-2,9.5388675E-3,-2.6785621E-2],"split_indices":[332,493,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.181076E0,2.022861E0,1.158215E0,1.0023845E0,1.0204765E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[3.9722314E-4,1.554788E-2,-1.2785774E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":288,"left_children":[1,-1,-1],"loss_changes":[1.1443629E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.7596493E0,1.554788E-2,-1.2785774E-2],"split_indices":[22,0,0],"split_type":[0,0,0],"sum_hessian":[3.1743214E0,1.352508E0,1.8218135E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[9.413297E-4,1.6712017E-2,-3.855807E-2,1.02269E-2,-2.760521E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false,false,false],"id":289,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.1171172E-2,0E0,1.4843634E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.2088545E-1,1.6712017E-2,-3.1847525E-1,1.02269E-2,-2.760521E-2],"split_indices":[301,0,147,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.1712408E0,1.1568984E0,2.0143425E0,1.0075169E0,1.0068257E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[4.891431E-4,1.5223842E-2,-1.2651781E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":290,"left_children":[1,-1,-1],"loss_changes":[1.1068191E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.6711326E0,1.5223842E-2,-1.2651781E-2],"split_indices":[149,0,0],"split_type":[0,0,0],"sum_hessian":[3.1641245E0,1.3657382E0,1.7983863E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[9.615068E-4,1.5102375E-2,-1.2658515E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":291,"left_children":[1,-1,-1],"loss_changes":[1.0991824E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.9723969E-1,1.5102375E-2,-1.2658515E-2],"split_indices":[95,0,0],"split_type":[0,0,0],"sum_hessian":[3.161157E0,1.3966407E0,1.7645162E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.2621633E-3,-1.302457E-2,1.468295E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":292,"left_children":[1,-1,-1],"loss_changes":[1.0982931E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.0041963E0,-1.302457E-2,1.468295E-2],"split_indices":[78,0,0],"split_type":[0,0,0],"sum_hessian":[3.1581962E0,1.6766394E0,1.4815568E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.2841789E-3,1.7653227E-2,-3.5676185E-2,-2.689333E-2,9.68395E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":293,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.0939268E-2,0E0,1.445704E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-2.8034773E0,1.7653227E-2,-7.708192E-2,-2.689333E-2,9.68395E-3],"split_indices":[139,0,84,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.1552083E0,1.0022604E0,2.152948E0,1.0220882E0,1.1308597E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[9.97543E-4,-1.5578915E-2,1.2658743E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":294,"left_children":[1,-1,-1],"loss_changes":[1.1236515E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.5833774E0,-1.5578915E-2,1.2658743E-2],"split_indices":[59,0,0],"split_type":[0,0,0],"sum_hessian":[3.1482468E0,1.2639595E0,1.8842872E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-6.3800726E-6,1.5798358E-2,-1.3461257E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":295,"left_children":[1,-1,-1],"loss_changes":[1.2157692E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.2516327E0,1.5798358E-2,-1.3461257E-2],"split_indices":[26,0,0],"split_type":[0,0,0],"sum_hessian":[3.1452143E0,1.36685E0,1.7783643E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[5.603155E-4,1.7039683E-2,-3.6944974E-2,-2.5591668E-2,9.1328835E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":296,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.0834811E-2,0E0,1.2669354E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.3116226E0,1.7039683E-2,1.1625337E-1,-2.5591668E-2,9.1328835E-3],"split_indices":[116,0,231,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.141942E0,1.0512234E0,2.0907187E0,1.062403E0,1.0283158E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[5.246849E-4,-1.1809985E-2,1.63887E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":297,"left_children":[1,-1,-1],"loss_changes":[1.1076245E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.959839E-1,-1.1809985E-2,1.63887E-2],"split_indices":[177,0,0],"split_type":[0,0,0],"sum_hessian":[3.1355338E0,1.9616193E0,1.1739146E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.4772285E-3,1.4557726E-2,-1.27532035E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":298,"left_children":[1,-1,-1],"loss_changes":[1.0615424E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[5.7764053E-2,1.4557726E-2,-1.27532035E-2],"split_indices":[121,0,0],"split_type":[0,0,0],"sum_hessian":[3.1325386E0,1.4637653E0,1.6687733E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.4721153E-3,1.4645153E-2,-1.2745657E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":299,"left_children":[1,-1,-1],"loss_changes":[1.0668616E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.350604E-1,1.4645153E-2,-1.2745657E-2],"split_indices":[260,0,0],"split_type":[0,0,0],"sum_hessian":[3.1296794E0,1.4535582E0,1.6761212E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.4923363E-3,1.2272802E-2,-1.5022424E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":300,"left_children":[1,-1,-1],"loss_changes":[1.0436716E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.0193767E-1,1.2272802E-2,-1.5022424E-2],"split_indices":[448,0,0],"split_type":[0,0,0],"sum_hessian":[3.1267877E0,1.889309E0,1.2374786E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[3.775975E-4,3.7905246E-2,-1.6109414E-2,2.5104353E-2,-8.183273E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,true,false,false,false],"id":301,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.0404394E-2,1.1295786E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[4.198103E-1,7.437801E-2,-1.6109414E-2,2.5104353E-2,-8.183273E-3],"split_indices":[452,164,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.1239958E0,2.0206966E0,1.103299E0,1.0203463E0,1.0003504E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[4.7048056E-4,1.7713403E-2,-3.7816714E-2,-2.7112672E-2,9.83211E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":302,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.146822E-2,0E0,1.4462901E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-3.1927748E0,1.7713403E-2,-6.960285E-2,-2.7112672E-2,9.83211E-3],"split_indices":[106,0,5,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.1180644E0,1.018202E0,2.0998626E0,1.0430026E0,1.05686E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[4.245156E-4,1.5690722E-2,-1.1984539E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":303,"left_children":[1,-1,-1],"loss_changes":[1.0699624E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.4897291E-1,1.5690722E-2,-1.1984539E-2],"split_indices":[198,0,0],"split_type":[0,0,0],"sum_hessian":[3.1109414E0,1.2321677E0,1.8787737E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.2010075E-3,1.6230466E-2,-1.1413192E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":304,"left_children":[1,-1,-1],"loss_changes":[1.058691E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.826052E-1,1.6230466E-2,-1.1413192E-2],"split_indices":[363,0,0],"split_type":[0,0,0],"sum_hessian":[3.1080635E0,1.1625017E0,1.9455618E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[2.0164317E-3,-1.3711398E-2,1.3393854E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":305,"left_children":[1,-1,-1],"loss_changes":[1.0391899E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.0247204E0,-1.3711398E-2,1.3393854E-2],"split_indices":[55,0,0],"split_type":[0,0,0],"sum_hessian":[3.1051984E0,1.4310759E0,1.6741225E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.334884E-3,1.2903406E-2,-1.3971047E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":306,"left_children":[1,-1,-1],"loss_changes":[1.0193504E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[7.9127884E-1,1.2903406E-2,-1.3971047E-2],"split_indices":[37,0,0],"split_type":[0,0,0],"sum_hessian":[3.1023824E0,1.7136731E0,1.3887093E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[6.7472405E-4,1.1667439E-2,-1.5389615E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":307,"left_children":[1,-1,-1],"loss_changes":[1.0137995E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.6356434E-1,1.1667439E-2,-1.5389615E-2],"split_indices":[216,0,0],"split_type":[0,0,0],"sum_hessian":[3.0996304E0,1.9312572E0,1.1683731E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-4.3731305E-4,-4.223986E-2,1.8269798E-2,9.605474E-3,-2.8443947E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":308,"left_children":[1,3,-1,-1,-1],"loss_changes":[1.307699E-2,1.4852691E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[2.9781587E0,-2.5591766E-2,1.8269798E-2,9.605474E-3,-2.8443947E-2],"split_indices":[16,176,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.0969176E0,2.0268853E0,1.0700325E0,1.0022367E0,1.0246485E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-4.5818303E-4,1.5509272E-2,-1.1753347E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":309,"left_children":[1,-1,-1],"loss_changes":[1.0283477E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.408786E0,1.5509272E-2,-1.1753347E-2],"split_indices":[80,0,0],"split_type":[0,0,0],"sum_hessian":[3.0892673E0,1.173446E0,1.9158212E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[5.997957E-4,-1.2268367E-2,1.4715558E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":310,"left_children":[1,-1,-1],"loss_changes":[1.0221795E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.3680219E-1,-1.2268367E-2,1.4715558E-2],"split_indices":[247,0,0],"split_type":[0,0,0],"sum_hessian":[3.0864933E0,1.7466451E0,1.3398482E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.020492E-3,-1.6137708E-2,3.7485696E-2,2.7423024E-2,-1.0278155E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,true,false,false],"id":311,"left_children":[1,-1,3,-1,-1],"loss_changes":[1.0178855E-2,0E0,1.4899401E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-4.357042E-1,-1.6137708E-2,6.6190815E-1,2.7423024E-2,-1.0278155E-2],"split_indices":[417,0,131,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.083733E0,1.0421104E0,2.0416226E0,1.0091051E0,1.0325176E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[9.7404054E-4,-1.500884E-2,1.3347995E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":312,"left_children":[1,-1,-1],"loss_changes":[1.1275269E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.6375225E0,-1.500884E-2,1.3347995E-2],"split_indices":[19,0,0],"split_type":[0,0,0],"sum_hessian":[3.0769272E0,1.3477751E0,1.7291521E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[2.7376178E-4,-1.3416635E-2,1.3650681E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":313,"left_children":[1,-1,-1],"loss_changes":[1.0325696E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.731128E-2,-1.3416635E-2,1.3650681E-2],"split_indices":[434,0,0],"split_type":[0,0,0],"sum_hessian":[3.0738807E0,1.5465157E0,1.527365E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[2.495536E-4,1.1684802E-2,-1.52103E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":314,"left_children":[1,-1,-1],"loss_changes":[1.00020645E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.2770464E-1,1.1684802E-2,-1.52103E-2],"split_indices":[281,0,0],"split_type":[0,0,0],"sum_hessian":[3.0710995E0,1.8792503E0,1.1918492E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-6.933675E-4,-1.4202144E-2,1.2533061E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":315,"left_children":[1,-1,-1],"loss_changes":[1.0037744E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.629901E-1,-1.4202144E-2,1.2533061E-2],"split_indices":[400,0,0],"split_type":[0,0,0],"sum_hessian":[3.0684223E0,1.4076539E0,1.6607684E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-8.967328E-4,1.1991304E-2,-1.477386E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":316,"left_children":[1,-1,-1],"loss_changes":[1.000197E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.8765771E-1,1.1991304E-2,-1.477386E-2],"split_indices":[224,0,0],"split_type":[0,0,0],"sum_hessian":[3.0656981E0,1.7553039E0,1.3103943E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.3140877E-3,-1.5024264E-2,1.1784417E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":317,"left_children":[1,-1,-1],"loss_changes":[1.0010815E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.3897004E0,-1.5024264E-2,1.1784417E-2],"split_indices":[36,0,0],"split_type":[0,0,0],"sum_hessian":[3.063001E0,1.285314E0,1.777687E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.7198066E-3,1.3888837E-2,-1.2356192E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":318,"left_children":[1,-1,-1],"loss_changes":[9.601324E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.111781E-1,1.3888837E-2,-1.2356192E-2],"split_indices":[326,0,0],"split_type":[0,0,0],"sum_hessian":[3.0602813E0,1.302567E0,1.7577143E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-8.2634063E-4,1.51120005E-2,-1.1746355E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":319,"left_children":[1,-1,-1],"loss_changes":[9.93514E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.2231853E0,1.51120005E-2,-1.1746355E-2],"split_indices":[25,0,0],"split_type":[0,0,0],"sum_hessian":[3.0577025E0,1.174506E0,1.8831966E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[2.52637E-4,-3.5148043E-2,1.6198233E-2,5.572422E-3,-2.1643223E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,true,false,false,false],"id":320,"left_children":[1,3,-1,-1,-1],"loss_changes":[9.612389E-3,7.402758E-3,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[3.9219052E-1,1.8908639E-2,1.6198233E-2,5.572422E-3,-2.1643223E-2],"split_indices":[442,274,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.0550137E0,2.0503695E0,1.0046442E0,1.0392203E0,1.0111492E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[2.0360258E-4,1.4209453E-2,-1.210164E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":321,"left_children":[1,-1,-1],"loss_changes":[9.655051E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.6307354E-1,1.4209453E-2,-1.210164E-2],"split_indices":[272,0,0],"split_type":[0,0,0],"sum_hessian":[3.0503602E0,1.3322878E0,1.7180724E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[6.6417834E-4,-1.5605827E-2,3.5626143E-2,2.5277145E-2,-9.374469E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,true,false,false],"id":322,"left_children":[1,-1,3,-1,-1],"loss_changes":[9.308906E-3,0E0,1.2475962E-2,0E0,0E0],"parents":[2147483647,0,0,2,2],"right_children":[2,-1,4,-1,-1],"split_conditions":[-4.0247116E0,-1.5605827E-2,7.3667336E-1,2.5277145E-2,-9.374469E-3],"split_indices":[94,0,64,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.0477576E0,1.0211371E0,2.0266204E0,1.0228631E0,1.0037572E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[5.4821034E-4,-1.494857E-2,1.1564498E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":323,"left_children":[1,-1,-1],"loss_changes":[9.658036E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.7680993E-1,-1.494857E-2,1.1564498E-2],"split_indices":[268,0,0],"split_type":[0,0,0],"sum_hessian":[3.0417564E0,1.1740471E0,1.8677093E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-4.550837E-4,1.1822784E-2,-1.4665976E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":324,"left_children":[1,-1,-1],"loss_changes":[9.724967E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.4121358E-1,1.1822784E-2,-1.4665976E-2],"split_indices":[302,0,0],"split_type":[0,0,0],"sum_hessian":[3.0391665E0,1.7692065E0,1.2699599E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.0241222E-3,-1.3883401E-2,1.23511525E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":325,"left_children":[1,-1,-1],"loss_changes":[9.612967E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.83142E0,-1.3883401E-2,1.23511525E-2],"split_indices":[31,0,0],"split_type":[0,0,0],"sum_hessian":[3.0365524E0,1.4184672E0,1.6180853E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.0978754E-3,-3.6068704E-2,1.570834E-2,1.0313833E-2,-2.6312327E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false,false,false],"id":326,"left_children":[1,3,-1,-1,-1],"loss_changes":[9.430104E-3,1.4040988E-2,0E0,0E0,0E0],"parents":[2147483647,0,0,1,1],"right_children":[2,4,-1,-1,-1],"split_conditions":[4.083275E-1,9.4299495E-2,1.570834E-2,1.0313833E-2,-2.6312327E-2],"split_indices":[348,212,0,0,0],"split_type":[0,0,0,0,0],"sum_hessian":[3.0339565E0,2.0307922E0,1.0031644E0,1.0003325E0,1.0304598E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"5","size_leaf_vector":"0"}},{"base_weights":[-8.5246074E-4,-1.1190161E-2,1.5267632E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":327,"left_children":[1,-1,-1],"loss_changes":[9.494203E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.2208195E0,-1.1190161E-2,1.5267632E-2],"split_indices":[88,0,0],"split_type":[0,0,0],"sum_hessian":[3.0275414E0,1.9401034E0,1.087438E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[4.175063E-4,-1.206036E-2,1.3911884E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":328,"left_children":[1,-1,-1],"loss_changes":[9.377445E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.5412903E0,-1.206036E-2,1.3911884E-2],"split_indices":[137,0,0],"split_type":[0,0,0],"sum_hessian":[3.0249617E0,1.6721815E0,1.3527802E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[7.53217E-4,-1.4790938E-2,1.1461747E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":329,"left_children":[1,-1,-1],"loss_changes":[9.4246585E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.236605E-1,-1.4790938E-2,1.1461747E-2],"split_indices":[366,0,0],"split_type":[0,0,0],"sum_hessian":[3.0224192E0,1.1581324E0,1.8642868E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-3.012048E-4,1.558493E-2,-1.0866028E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":330,"left_children":[1,-1,-1],"loss_changes":[9.426126E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.947982E0,1.558493E-2,-1.0866028E-2],"split_indices":[162,0,0],"split_type":[0,0,0],"sum_hessian":[3.019882E0,1.04843E0,1.971452E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[9.6480286E-4,1.4226293E-2,-1.176148E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":331,"left_children":[1,-1,-1],"loss_changes":[9.355987E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.2528253E0,1.4226293E-2,-1.176148E-2],"split_indices":[72,0,0],"split_type":[0,0,0],"sum_hessian":[3.0173204E0,1.3154693E0,1.7018511E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.2909314E-3,-1.1745599E-2,1.4218394E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":332,"left_children":[1,-1,-1],"loss_changes":[9.341486E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[5.617125E-1,-1.1745599E-2,1.4218394E-2],"split_indices":[158,0,0],"split_type":[0,0,0],"sum_hessian":[3.0148096E0,1.6863236E0,1.328486E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.5216304E-3,-1.4139364E-2,1.1862355E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":333,"left_children":[1,-1,-1],"loss_changes":[9.2853885E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.6732554E0,-1.4139364E-2,1.1862355E-2],"split_indices":[71,0,0],"split_type":[0,0,0],"sum_hessian":[3.0122871E0,1.2162368E0,1.7960503E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[4.963228E-4,1.1556556E-2,-1.4268303E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":334,"left_children":[1,-1,-1],"loss_changes":[9.159612E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.9180316E-1,1.1556556E-2,-1.4268303E-2],"split_indices":[258,0,0],"split_type":[0,0,0],"sum_hessian":[3.0097685E0,1.7910292E0,1.2187393E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-3.290114E-4,-1.4484109E-2,1.14363665E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":335,"left_children":[1,-1,-1],"loss_changes":[9.228901E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.0963696E-1,-1.4484109E-2,1.14363665E-2],"split_indices":[454,0,0],"split_type":[0,0,0],"sum_hessian":[3.0072856E0,1.2245227E0,1.7827629E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-9.928169E-4,1.3007537E-2,-1.2517014E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":336,"left_children":[1,-1,-1],"loss_changes":[9.043512E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[5.7764053E-2,1.3007537E-2,-1.2517014E-2],"split_indices":[121,0,0],"split_type":[0,0,0],"sum_hessian":[3.004784E0,1.4075701E0,1.597214E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-5.731718E-4,1.243546E-2,-1.3198931E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":337,"left_children":[1,-1,-1],"loss_changes":[9.127394E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[9.095875E-2,1.243546E-2,-1.3198931E-2],"split_indices":[189,0,0],"split_type":[0,0,0],"sum_hessian":[3.0023427E0,1.5488168E0,1.4535259E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-5.9238507E-4,1.4874929E-2,-1.0992016E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":338,"left_children":[1,-1,-1],"loss_changes":[9.051351E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.0583322E-1,1.4874929E-2,-1.0992016E-2],"split_indices":[395,0,0],"split_type":[0,0,0],"sum_hessian":[2.9998827E0,1.097192E0,1.9026906E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[5.574434E-4,1.504328E-2,-1.0731629E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":339,"left_children":[1,-1,-1],"loss_changes":[8.994998E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.1407532E-1,1.504328E-2,-1.0731629E-2],"split_indices":[494,0,0],"split_type":[0,0,0],"sum_hessian":[2.9974282E0,1.106663E0,1.8907652E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.4710046E-3,1.1047189E-2,-1.4711736E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":340,"left_children":[1,-1,-1],"loss_changes":[8.939643E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.810327E-1,1.1047189E-2,-1.4711736E-2],"split_indices":[238,0,0],"split_type":[0,0,0],"sum_hessian":[2.9950156E0,1.9212538E0,1.0737618E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.11397225E-4,1.3665918E-2,-1.181533E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":341,"left_children":[1,-1,-1],"loss_changes":[8.959776E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.7238207E-1,1.3665918E-2,-1.181533E-2],"split_indices":[428,0,0],"split_type":[0,0,0],"sum_hessian":[2.9925723E0,1.3202285E0,1.6723439E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[5.422518E-4,-1.0577791E-2,1.4941744E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":342,"left_children":[1,-1,-1],"loss_changes":[8.793603E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.4950408E-1,-1.0577791E-2,1.4941744E-2],"split_indices":[332,0,0],"split_type":[0,0,0],"sum_hessian":[2.9901555E0,1.8963115E0,1.0938439E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.4759309E-3,1.2867292E-2,-1.2395411E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":343,"left_children":[1,-1,-1],"loss_changes":[8.839776E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[7.407241E-2,1.2867292E-2,-1.2395411E-2],"split_indices":[325,0,0],"split_type":[0,0,0],"sum_hessian":[2.9877844E0,1.5172026E0,1.4705818E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.1518268E-3,-1.3314571E-2,1.2116731E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":344,"left_children":[1,-1,-1],"loss_changes":[8.912952E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-7.8551006E-1,-1.3314571E-2,1.2116731E-2],"split_indices":[11,0,0],"split_type":[0,0,0],"sum_hessian":[2.9854069E0,1.3211429E0,1.664264E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[5.062641E-4,1.4962775E-2,-1.0837386E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":345,"left_children":[1,-1,-1],"loss_changes":[9.004792E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.64045E-1,1.4962775E-2,-1.0837386E-2],"split_indices":[390,0,0],"split_type":[0,0,0],"sum_hessian":[2.9829843E0,1.116555E0,1.8664293E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.392705E-3,1.1440957E-2,-1.4214883E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":346,"left_children":[1,-1,-1],"loss_changes":[8.940998E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.1681415E-1,1.1440957E-2,-1.4214883E-2],"split_indices":[333,0,0],"split_type":[0,0,0],"sum_hessian":[2.9805627E0,1.8243566E0,1.1562061E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[2.7694678E-4,1.1489158E-2,-1.3887252E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":347,"left_children":[1,-1,-1],"loss_changes":[8.816251E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.0457262E-1,1.1489158E-2,-1.3887252E-2],"split_indices":[168,0,0],"split_type":[0,0,0],"sum_hessian":[2.9781806E0,1.7373365E0,1.2408441E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-4.1013473E-4,-1.2738989E-2,1.6028484E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":348,"left_children":[1,-1,-1],"loss_changes":[1.1270261E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.7952664E0,-1.2738989E-2,1.6028484E-2],"split_indices":[133,0,0],"split_type":[0,0,0],"sum_hessian":[2.9758043E0,1.789393E0,1.1864114E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[5.4209837E-4,1.4207229E-2,-1.1217104E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":349,"left_children":[1,-1,-1],"loss_changes":[8.825566E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.248637E0,1.4207229E-2,-1.1217104E-2],"split_indices":[86,0,0],"split_type":[0,0,0],"sum_hessian":[2.9727309E0,1.2193592E0,1.7533717E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.1341351E-3,-1.4314602E-2,1.1015584E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":350,"left_children":[1,-1,-1],"loss_changes":[8.65357E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.332079E-1,-1.4314602E-2,1.1015584E-2],"split_indices":[294,0,0],"split_type":[0,0,0],"sum_hessian":[2.9703135E0,1.1081583E0,1.8621552E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-3.794957E-5,1.4921607E-2,-1.0314466E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":351,"left_children":[1,-1,-1],"loss_changes":[8.493417E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-4.1991648E-1,1.4921607E-2,-1.0314466E-2],"split_indices":[210,0,0],"split_type":[0,0,0],"sum_hessian":[2.9679947E0,1.0287244E0,1.9392703E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.1364982E-3,1.16879195E-2,-1.3235832E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":352,"left_children":[1,-1,-1],"loss_changes":[8.507071E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.3402748E-1,1.16879195E-2,-1.3235832E-2],"split_indices":[416,0,0],"split_type":[0,0,0],"sum_hessian":[2.9657001E0,1.6912992E0,1.274401E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[4.0546057E-4,-1.1933194E-2,1.3025174E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":353,"left_children":[1,-1,-1],"loss_changes":[8.577127E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.0041963E0,-1.1933194E-2,1.3025174E-2],"split_indices":[78,0,0],"split_type":[0,0,0],"sum_hessian":[2.963419E0,1.5709732E0,1.3924458E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[5.593826E-4,1.1795248E-2,-1.3143553E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":354,"left_children":[1,-1,-1],"loss_changes":[8.534606E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[7.9127884E-1,1.1795248E-2,-1.3143553E-2],"split_indices":[37,0,0],"split_type":[0,0,0],"sum_hessian":[2.9610727E0,1.6412997E0,1.319773E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[5.088446E-5,1.1631351E-2,-1.3208367E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":355,"left_children":[1,-1,-1],"loss_changes":[8.463602E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.6349456E-1,1.1631351E-2,-1.3208367E-2],"split_indices":[269,0,0],"split_type":[0,0,0],"sum_hessian":[2.9587624E0,1.6392242E0,1.3195382E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-3.634518E-4,-1.3802984E-2,1.1040242E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":356,"left_children":[1,-1,-1],"loss_changes":[8.405065E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.7813606E0,-1.3802984E-2,1.1040242E-2],"split_indices":[10,0,0],"split_type":[0,0,0],"sum_hessian":[2.9564908E0,1.220012E0,1.7364788E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-9.539414E-4,1.1041763E-2,-1.3627306E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":357,"left_children":[1,-1,-1],"loss_changes":[8.311791E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[7.566936E-1,1.1041763E-2,-1.3627306E-2],"split_indices":[130,0,0],"split_type":[0,0,0],"sum_hessian":[2.9542227E0,1.6908627E0,1.26336E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.3228555E-3,1.3862761E-2,-1.1015907E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":358,"left_children":[1,-1,-1],"loss_changes":[8.345947E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.6847544E-1,1.3862761E-2,-1.1015907E-2],"split_indices":[341,0,0],"split_type":[0,0,0],"sum_hessian":[2.9519892E0,1.1296268E0,1.8223624E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.0503286E-4,1.3003037E-2,-1.1520276E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":359,"left_children":[1,-1,-1],"loss_changes":[8.234308E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.058481E0,1.3003037E-2,-1.1520276E-2],"split_indices":[29,0,0],"split_type":[0,0,0],"sum_hessian":[2.9497328E0,1.315321E0,1.6344118E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[2.3206978E-4,1.2927034E-2,-1.1400411E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":360,"left_children":[1,-1,-1],"loss_changes":[8.105936E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.276636E-1,1.2927034E-2,-1.1400411E-2],"split_indices":[329,0,0],"split_type":[0,0,0],"sum_hessian":[2.947524E0,1.3298229E0,1.6177012E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[5.498366E-4,-1.0347498E-2,1.4370226E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":361,"left_children":[1,-1,-1],"loss_changes":[8.198469E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.959839E-1,-1.0347498E-2,1.4370226E-2],"split_indices":[177,0,0],"split_type":[0,0,0],"sum_hessian":[2.94534E0,1.84876E0,1.0965799E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.4071995E-3,1.2436095E-2,-1.2119123E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":362,"left_children":[1,-1,-1],"loss_changes":[8.275824E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.1725941E-2,1.2436095E-2,-1.2119123E-2],"split_indices":[491,0,0],"split_type":[0,0,0],"sum_hessian":[2.9431148E0,1.5074441E0,1.4356706E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.0659925E-3,-1.2924983E-2,1.1716375E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":363,"left_children":[1,-1,-1],"loss_changes":[8.292131E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.8090875E-1,-1.2924983E-2,1.1716375E-2],"split_indices":[413,0,0],"split_type":[0,0,0],"sum_hessian":[2.9408867E0,1.298128E0,1.6427587E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[4.392538E-4,1.11752385E-2,-1.3412329E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":364,"left_children":[1,-1,-1],"loss_changes":[8.211175E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.246585E-1,1.11752385E-2,-1.3412329E-2],"split_indices":[226,0,0],"split_type":[0,0,0],"sum_hessian":[2.9386556E0,1.7151078E0,1.2235478E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.610251E-4,-1.2640188E-2,1.1924879E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":365,"left_children":[1,-1,-1],"loss_changes":[8.269734E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.157902E0,-1.2640188E-2,1.1924879E-2],"split_indices":[108,0,0],"split_type":[0,0,0],"sum_hessian":[2.9364183E0,1.4088857E0,1.5275326E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-3.6520493E-4,-1.3521551E-2,1.1029215E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":366,"left_children":[1,-1,-1],"loss_changes":[8.187459E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.9666393E0,-1.3521551E-2,1.1029215E-2],"split_indices":[151,0,0],"split_type":[0,0,0],"sum_hessian":[2.9341776E0,1.2341926E0,1.699985E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-8.9025655E-4,1.2938599E-2,-1.145645E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":367,"left_children":[1,-1,-1],"loss_changes":[8.102552E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.6711326E0,1.2938599E-2,-1.145645E-2],"split_indices":[149,0,0],"split_type":[0,0,0],"sum_hessian":[2.931961E0,1.2731099E0,1.6588511E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.46041E-4,1.1484201E-2,-1.2897748E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":368,"left_children":[1,-1,-1],"loss_changes":[8.1176525E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[4.4586396E-1,1.1484201E-2,-1.2897748E-2],"split_indices":[136,0,0],"split_type":[0,0,0],"sum_hessian":[2.9297764E0,1.5958936E0,1.3338828E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-5.322152E-4,-1.3774195E-2,1.0632826E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":369,"left_children":[1,-1,-1],"loss_changes":[8.039509E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.2397411E-1,-1.3774195E-2,1.0632826E-2],"split_indices":[436,0,0],"split_type":[0,0,0],"sum_hessian":[2.927577E0,1.1723734E0,1.7552036E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.1741284E-3,1.1644055E-2,-1.2578327E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":370,"left_children":[1,-1,-1],"loss_changes":[8.024364E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[5.8762825E-1,1.1644055E-2,-1.2578327E-2],"split_indices":[132,0,0],"split_type":[0,0,0],"sum_hessian":[2.925404E0,1.5006073E0,1.4247968E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.0612567E-3,1.3910533E-2,-1.0553927E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":371,"left_children":[1,-1,-1],"loss_changes":[7.979871E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.2088545E-1,1.3910533E-2,-1.0553927E-2],"split_indices":[301,0,0],"split_type":[0,0,0],"sum_hessian":[2.9231997E0,1.0728041E0,1.8503956E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.1401064E-4,1.1553772E-2,-1.266373E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":372,"left_children":[1,-1,-1],"loss_changes":[7.998478E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[7.3668994E-2,1.1553772E-2,-1.266373E-2],"split_indices":[233,0,0],"split_type":[0,0,0],"sum_hessian":[2.9210413E0,1.5788313E0,1.3422099E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.0395094E-4,-1.2720208E-2,1.1426867E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":373,"left_children":[1,-1,-1],"loss_changes":[7.947411E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-8.7341E-2,-1.2720208E-2,1.1426867E-2],"split_indices":[298,0,0],"split_type":[0,0,0],"sum_hessian":[2.9188933E0,1.3376462E0,1.5812471E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-4.7351478E-4,1.3640482E-2,-1.1194299E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":374,"left_children":[1,-1,-1],"loss_changes":[8.325831E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.106997E0,1.3640482E-2,-1.1194299E-2],"split_indices":[22,0,0],"split_type":[0,0,0],"sum_hessian":[2.9167411E0,1.1938217E0,1.7229195E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[2.9419555E-4,-1.2820826E-2,1.1227978E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":375,"left_children":[1,-1,-1],"loss_changes":[7.854094E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.5198672E-1,-1.2820826E-2,1.1227978E-2],"split_indices":[495,0,0],"split_type":[0,0,0],"sum_hessian":[2.914482E0,1.2801223E0,1.6343598E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.0157002E-4,-1.2709298E-2,1.14904065E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":376,"left_children":[1,-1,-1],"loss_changes":[7.9739215E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-4.559751E-2,-1.2709298E-2,1.14904065E-2],"split_indices":[182,0,0],"split_type":[0,0,0],"sum_hessian":[2.9123769E0,1.3422513E0,1.5701256E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-4.5335168E-4,9.849562E-3,-1.4509116E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":377,"left_children":[1,-1,-1],"loss_changes":[7.823511E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.4309313E-1,9.849562E-3,-1.4509116E-2],"split_indices":[193,0,0],"split_type":[0,0,0],"sum_hessian":[2.9102092E0,1.9029073E0,1.0073019E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.5111398E-3,-1.3956255E-2,1.0625386E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":378,"left_children":[1,-1,-1],"loss_changes":[8.143572E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.968687E-1,-1.3956255E-2,1.0625386E-2],"split_indices":[488,0,0],"split_type":[0,0,0],"sum_hessian":[2.9080887E0,1.1935898E0,1.7144989E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.9123702E-3,1.3692961E-2,-1.0724542E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":379,"left_children":[1,-1,-1],"loss_changes":[7.9166265E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.4197664E0,1.3692961E-2,-1.0724542E-2],"split_indices":[25,0,0],"split_type":[0,0,0],"sum_hessian":[2.9058917E0,1.0629704E0,1.8429213E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-5.780914E-4,1.2902128E-2,-1.1062684E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":380,"left_children":[1,-1,-1],"loss_changes":[7.761756E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.111781E-1,1.2902128E-2,-1.1062684E-2],"split_indices":[326,0,0],"split_type":[0,0,0],"sum_hessian":[2.9037364E0,1.235422E0,1.6683143E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[6.7318586E-5,1.0431418E-2,-1.37291E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":381,"left_children":[1,-1,-1],"loss_changes":[7.796924E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.0983895E-1,1.0431418E-2,-1.37291E-2],"split_indices":[205,0,0],"split_type":[0,0,0],"sum_hessian":[2.9016485E0,1.7885998E0,1.1130487E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-7.9984276E-4,1.2450418E-2,-1.1292434E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":382,"left_children":[1,-1,-1],"loss_changes":[7.6394095E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.9723969E-1,1.2450418E-2,-1.1292434E-2],"split_indices":[95,0,0],"split_type":[0,0,0],"sum_hessian":[2.8995504E0,1.290885E0,1.6086655E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.6251556E-4,1.0093915E-2,-1.37399705E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":383,"left_children":[1,-1,-1],"loss_changes":[7.5591924E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.622104E-1,1.0093915E-2,-1.37399705E-2],"split_indices":[349,0,0],"split_type":[0,0,0],"sum_hessian":[2.8974836E0,1.8104664E0,1.0870172E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.1207209E-3,-1.2573426E-2,1.1224452E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":384,"left_children":[1,-1,-1],"loss_changes":[7.6913317E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-5.3432945E-2,-1.2573426E-2,1.1224452E-2],"split_indices":[383,0,0],"split_type":[0,0,0],"sum_hessian":[2.895446E0,1.3640097E0,1.5314363E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.1319283E-3,-1.1544722E-2,1.2179288E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":385,"left_children":[1,-1,-1],"loss_changes":[7.630562E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.1221077E-1,-1.1544722E-2,1.2179288E-2],"split_indices":[76,0,0],"split_type":[0,0,0],"sum_hessian":[2.8933625E0,1.5678532E0,1.3255093E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-6.294205E-4,-9.987488E-3,1.3580227E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":386,"left_children":[1,-1,-1],"loss_changes":[7.3404037E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.2208195E0,-9.987488E-3,1.3580227E-2],"split_indices":[88,0,0],"split_type":[0,0,0],"sum_hessian":[2.8912745E0,1.8496352E0,1.0416392E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[4.8350365E-4,1.433484E-2,-1.0167829E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":387,"left_children":[1,-1,-1],"loss_changes":[7.94334E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.8274283E0,1.433484E-2,-1.0167829E-2],"split_indices":[88,0,0],"split_type":[0,0,0],"sum_hessian":[2.88927E0,1.0519154E0,1.8373547E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.4116698E-3,-1.39799435E-2,1.0390536E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":388,"left_children":[1,-1,-1],"loss_changes":[7.814394E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-4.6706706E-1,-1.39799435E-2,1.0390536E-2],"split_indices":[323,0,0],"split_type":[0,0,0],"sum_hessian":[2.8871403E0,1.0161195E0,1.8710208E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[6.365697E-5,1.0663772E-2,-1.289736E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":389,"left_children":[1,-1,-1],"loss_changes":[7.463227E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.5006947E-1,1.0663772E-2,-1.289736E-2],"split_indices":[275,0,0],"split_type":[0,0,0],"sum_hessian":[2.8850117E0,1.6772037E0,1.207808E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-5.310918E-4,-1.0871936E-2,1.2822543E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":390,"left_children":[1,-1,-1],"loss_changes":[7.5490223E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.8012346E-1,-1.0871936E-2,1.2822543E-2],"split_indices":[429,0,0],"split_type":[0,0,0],"sum_hessian":[2.8829894E0,1.6685967E0,1.2143927E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.3106722E-4,1.2818971E-2,-1.0964047E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":391,"left_children":[1,-1,-1],"loss_changes":[7.625384E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.2528253E0,1.2818971E-2,-1.0964047E-2],"split_indices":[72,0,0],"split_type":[0,0,0],"sum_hessian":[2.8809552E0,1.2565523E0,1.6244029E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[5.702898E-4,1.0055756E-2,-1.3924153E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":392,"left_children":[1,-1,-1],"loss_changes":[7.560587E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[4.198103E-1,1.0055756E-2,-1.3924153E-2],"split_indices":[452,0,0],"split_type":[0,0,0],"sum_hessian":[2.878903E0,1.8606538E0,1.0182492E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-5.9894455E-4,1.32027315E-2,-1.0306517E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":393,"left_children":[1,-1,-1],"loss_changes":[7.349678E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.3680513E-1,1.32027315E-2,-1.0306517E-2],"split_indices":[288,0,0],"split_type":[0,0,0],"sum_hessian":[2.876853E0,1.108394E0,1.768459E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[3.281859E-4,-1.2943274E-2,1.0562175E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":394,"left_children":[1,-1,-1],"loss_changes":[7.3943622E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.0963696E-1,-1.2943274E-2,1.0562175E-2],"split_indices":[454,0,0],"split_type":[0,0,0],"sum_hessian":[2.8748655E0,1.1742907E0,1.7005749E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-3.8524825E-4,1.121822E-2,-1.4957073E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":395,"left_children":[1,-1,-1],"loss_changes":[9.102743E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.935821E-1,1.121822E-2,-1.4957073E-2],"split_indices":[314,0,0],"split_type":[0,0,0],"sum_hessian":[2.872857E0,1.7673453E0,1.1055118E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.2370952E-3,1.3625764E-2,-1.0385364E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":396,"left_children":[1,-1,-1],"loss_changes":[7.6001603E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.0583322E-1,1.3625764E-2,-1.0385364E-2],"split_indices":[395,0,0],"split_type":[0,0,0],"sum_hessian":[2.870393E0,1.0467339E0,1.8236592E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.8027613E-5,1.1954138E-2,-1.1396398E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":397,"left_children":[1,-1,-1],"loss_changes":[7.3690386E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.1685818E-3,1.1954138E-2,-1.1396398E-2],"split_indices":[189,0,0],"split_type":[0,0,0],"sum_hessian":[2.868325E0,1.3746283E0,1.4936967E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.2453824E-4,1.33664515E-2,-1.0023629E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":398,"left_children":[1,-1,-1],"loss_changes":[7.2496766E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.408786E0,1.33664515E-2,-1.0023629E-2],"split_indices":[80,0,0],"split_type":[0,0,0],"sum_hessian":[2.866337E0,1.091605E0,1.7747321E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[9.534003E-4,-9.6742045E-3,1.3742328E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":399,"left_children":[1,-1,-1],"loss_changes":[7.2319945E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.081163E0,-9.6742045E-3,1.3742328E-2],"split_indices":[100,0,0],"split_type":[0,0,0],"sum_hessian":[2.8643723E0,1.8075255E0,1.0568467E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.7222081E-3,1.253872E-2,-1.0745793E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":400,"left_children":[1,-1,-1],"loss_changes":[7.307791E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-5.4437328E-2,1.253872E-2,-1.0745793E-2],"split_indices":[180,0,0],"split_type":[0,0,0],"sum_hessian":[2.8624148E0,1.329706E0,1.5327089E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.6711816E-3,-1.1523234E-2,1.1676184E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":401,"left_children":[1,-1,-1],"loss_changes":[7.258736E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-5.077629E-1,-1.1523234E-2,1.1676184E-2],"split_indices":[47,0,0],"split_type":[0,0,0],"sum_hessian":[2.8604474E0,1.362819E0,1.4976285E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.209896E-3,1.0136111E-2,-1.3255841E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":402,"left_children":[1,-1,-1],"loss_changes":[7.1990998E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.5597008E-1,1.0136111E-2,-1.3255841E-2],"split_indices":[303,0,0],"split_type":[0,0,0],"sum_hessian":[2.8584661E0,1.8130853E0,1.0453808E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[3.831482E-5,-1.1062499E-2,1.1869284E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":403,"left_children":[1,-1,-1],"loss_changes":[7.085724E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.0041963E0,-1.1062499E-2,1.1869284E-2],"split_indices":[78,0,0],"split_type":[0,0,0],"sum_hessian":[2.8565176E0,1.5117565E0,1.344761E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[2.3587827E-4,-1.2438689E-2,1.0504883E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":404,"left_children":[1,-1,-1],"loss_changes":[7.042063E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.371829E-1,-1.2438689E-2,1.0504883E-2],"split_indices":[230,0,0],"split_type":[0,0,0],"sum_hessian":[2.8545737E0,1.2108139E0,1.6437598E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-3.3716165E-4,1.0319551E-2,-1.2694488E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":405,"left_children":[1,-1,-1],"loss_changes":[7.0732594E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.4121358E-1,1.0319551E-2,-1.2694488E-2],"split_indices":[302,0,0],"split_type":[0,0,0],"sum_hessian":[2.852675E0,1.6597909E0,1.1928841E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-8.532565E-4,-1.2188359E-2,1.0842381E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":406,"left_children":[1,-1,-1],"loss_changes":[7.1345256E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.462903E-1,-1.2188359E-2,1.0842381E-2],"split_indices":[400,0,0],"split_type":[0,0,0],"sum_hessian":[2.8507671E0,1.3264372E0,1.5243299E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-9.4797794E-4,1.1750175E-2,-1.133607E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":407,"left_children":[1,-1,-1],"loss_changes":[7.1677947E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[5.7764053E-2,1.1750175E-2,-1.133607E-2],"split_indices":[121,0,0],"split_type":[0,0,0],"sum_hessian":[2.8488164E0,1.3335083E0,1.5153081E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-5.561953E-4,1.0049435E-2,-1.3323367E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":408,"left_children":[1,-1,-1],"loss_changes":[7.232818E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.2770464E-1,1.0049435E-2,-1.3323367E-2],"split_indices":[281,0,0],"split_type":[0,0,0],"sum_hessian":[2.8468754E0,1.7354364E0,1.111439E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.2419467E-3,-1.3349274E-2,9.6524265E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":409,"left_children":[1,-1,-1],"loss_changes":[6.9893952E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.7680993E-1,-1.3349274E-2,9.6524265E-3],"split_indices":[268,0,0],"split_type":[0,0,0],"sum_hessian":[2.8449314E0,1.0954062E0,1.7495252E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.8328431E-3,1.0172123E-2,-1.2536831E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":410,"left_children":[1,-1,-1],"loss_changes":[6.9050053E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[7.9127884E-1,1.0172123E-2,-1.2536831E-2],"split_indices":[37,0,0],"split_type":[0,0,0],"sum_hessian":[2.8430493E0,1.5806288E0,1.2624205E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.9000582E-3,1.1584969E-2,-1.1189373E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":411,"left_children":[1,-1,-1],"loss_changes":[6.949328E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.5102413E-1,1.1584969E-2,-1.1189373E-2],"split_indices":[428,0,0],"split_type":[0,0,0],"sum_hessian":[2.8411694E0,1.282398E0,1.5587714E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.2275357E-3,-1.3359271E-2,1.1683097E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":412,"left_children":[1,-1,-1],"loss_changes":[8.412819E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.0247204E0,-1.3359271E-2,1.1683097E-2],"split_indices":[55,0,0],"split_type":[0,0,0],"sum_hessian":[2.839287E0,1.314147E0,1.52514E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.2974282E-3,-1.0978335E-2,1.1964733E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":413,"left_children":[1,-1,-1],"loss_changes":[7.036657E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[5.617125E-1,-1.0978335E-2,1.1964733E-2],"split_indices":[158,0,0],"split_type":[0,0,0],"sum_hessian":[2.8370001E0,1.587574E0,1.2494261E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-6.508897E-4,-1.1567906E-2,1.1355145E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":414,"left_children":[1,-1,-1],"loss_changes":[7.0569753E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.731128E-2,-1.1567906E-2,1.1355145E-2],"split_indices":[434,0,0],"split_type":[0,0,0],"sum_hessian":[2.8350923E0,1.4277763E0,1.407316E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-5.1175256E-4,-1.3342233E-2,1.0489346E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":415,"left_children":[1,-1,-1],"loss_changes":[7.533307E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.6732554E0,-1.3342233E-2,1.0489346E-2],"split_indices":[71,0,0],"split_type":[0,0,0],"sum_hessian":[2.8331752E0,1.1519908E0,1.6811844E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.1039246E-3,1.0018935E-2,-1.3586317E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":416,"left_children":[1,-1,-1],"loss_changes":[7.3524555E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.1681415E-1,1.0018935E-2,-1.3586317E-2],"split_indices":[333,0,0],"split_type":[0,0,0],"sum_hessian":[2.8311262E0,1.7268689E0,1.1042573E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.7054108E-3,9.718118E-3,-1.3117838E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":417,"left_children":[1,-1,-1],"loss_changes":[6.9031464E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.9180316E-1,9.718118E-3,-1.3117838E-2],"split_indices":[258,0,0],"split_type":[0,0,0],"sum_hessian":[2.8291643E0,1.6882651E0,1.1408992E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.0837628E-3,1.40626645E-2,-1.5942525E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":418,"left_children":[1,-1,-1],"loss_changes":[1.2058334E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.1239257E-1,1.40626645E-2,-1.5942525E-2],"split_indices":[147,0,0],"split_type":[0,0,0],"sum_hessian":[2.8272758E0,1.4851176E0,1.3421582E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.9526188E-3,1.0499198E-2,-1.2624044E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":419,"left_children":[1,-1,-1],"loss_changes":[7.142579E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.21507026E-1,1.0499198E-2,-1.2624044E-2],"split_indices":[186,0,0],"split_type":[0,0,0],"sum_hessian":[2.8240025E0,1.5367712E0,1.2872313E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.9236548E-3,-1.3862189E-2,1.0584934E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":420,"left_children":[1,-1,-1],"loss_changes":[7.927796E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.3897004E0,-1.3862189E-2,1.0584934E-2],"split_indices":[36,0,0],"split_type":[0,0,0],"sum_hessian":[2.8220828E0,1.1780534E0,1.6440294E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.2064997E-3,-1.1831901E-2,1.0985622E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":421,"left_children":[1,-1,-1],"loss_changes":[6.966252E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[6.0760444E-1,-1.1831901E-2,1.0985622E-2],"split_indices":[11,0,0],"split_type":[0,0,0],"sum_hessian":[2.8199172E0,1.4313939E0,1.3885233E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.7675705E-3,-1.3408846E-2,9.477742E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":422,"left_children":[1,-1,-1],"loss_changes":[6.8798414E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.8923833E-1,-1.3408846E-2,9.477742E-3],"split_indices":[396,0,0],"split_type":[0,0,0],"sum_hessian":[2.8180168E0,1.0836877E0,1.7343291E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.2693146E-3,-1.3044895E-2,9.670705E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":423,"left_children":[1,-1,-1],"loss_changes":[6.8285833E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.0657914E0,-1.3044895E-2,9.670705E-3],"split_indices":[94,0,0],"split_type":[0,0,0],"sum_hessian":[2.816163E0,1.1647553E0,1.6514077E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.4747737E-3,1.2717916E-2,-1.0408244E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":424,"left_children":[1,-1,-1],"loss_changes":[6.9847773E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.1407532E-1,1.2717916E-2,-1.0408244E-2],"split_indices":[494,0,0],"split_type":[0,0,0],"sum_hessian":[2.8142896E0,1.0442845E0,1.7700051E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.1269113E-3,-9.89083E-3,1.3028991E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":425,"left_children":[1,-1,-1],"loss_changes":[6.840917E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.4950408E-1,-9.89083E-3,1.3028991E-2],"split_indices":[332,0,0],"split_type":[0,0,0],"sum_hessian":[2.812406E0,1.7918922E0,1.0205139E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[3.523503E-5,9.586627E-3,-1.3357067E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":426,"left_children":[1,-1,-1],"loss_changes":[6.842616E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.810327E-1,9.586627E-3,-1.3357067E-2],"split_indices":[238,0,0],"split_type":[0,0,0],"sum_hessian":[2.810556E0,1.8023037E0,1.0082523E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-9.66216E-4,-1.1657339E-2,1.07154185E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":427,"left_children":[1,-1,-1],"loss_changes":[6.682081E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.157902E0,-1.1657339E-2,1.07154185E-2],"split_indices":[108,0,0],"split_type":[0,0,0],"sum_hessian":[2.8086762E0,1.3524584E0,1.4562179E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-9.2451164E-4,-1.0067565E-2,1.254058E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":428,"left_children":[1,-1,-1],"loss_changes":[6.7108558E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.9268E-1,-1.0067565E-2,1.254058E-2],"split_indices":[363,0,0],"split_type":[0,0,0],"sum_hessian":[2.8068602E0,1.7130337E0,1.0938265E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[2.6106038E-6,1.3144574E-2,-9.817142E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":429,"left_children":[1,-1,-1],"loss_changes":[6.8896036E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.826052E-1,1.3144574E-2,-9.817142E-3],"split_indices":[363,0,0],"split_type":[0,0,0],"sum_hessian":[2.8050532E0,1.0545012E0,1.750552E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[8.7776745E-4,9.689233E-3,-1.2608515E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":430,"left_children":[1,-1,-1],"loss_changes":[6.4844624E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.2757968E-1,9.689233E-3,-1.2608515E-2],"split_indices":[267,0,0],"split_type":[0,0,0],"sum_hessian":[2.8031774E0,1.7609257E0,1.0422517E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.5387895E-4,1.0896012E-2,-1.1276562E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":431,"left_children":[1,-1,-1],"loss_changes":[6.5556294E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[7.5094607E-3,1.0896012E-2,-1.1276562E-2],"split_indices":[226,0,0],"split_type":[0,0,0],"sum_hessian":[2.8014174E0,1.4339976E0,1.3674197E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.0779704E-4,-1.0402737E-2,1.1709108E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":432,"left_children":[1,-1,-1],"loss_changes":[6.4922585E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.5412903E0,-1.0402737E-2,1.1709108E-2],"split_indices":[137,0,0],"split_type":[0,0,0],"sum_hessian":[2.7996306E0,1.5523089E0,1.2473217E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.9864379E-4,-1.1591203E-2,1.0526551E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":433,"left_children":[1,-1,-1],"loss_changes":[6.5017603E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-7.708192E-2,-1.1591203E-2,1.0526551E-2],"split_indices":[84,0,0],"split_type":[0,0,0],"sum_hessian":[2.7978613E0,1.2732239E0,1.5246375E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.4182448E-4,-1.2984489E-2,9.459193E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":434,"left_children":[1,-1,-1],"loss_changes":[6.5514673E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.4296675E0,-1.2984489E-2,9.459193E-3],"split_indices":[59,0,0],"split_type":[0,0,0],"sum_hessian":[2.7960935E0,1.0285747E0,1.7675188E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.0294764E-3,1.1646142E-2,-1.0479409E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":435,"left_children":[1,-1,-1],"loss_changes":[6.482117E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.6307354E-1,1.1646142E-2,-1.0479409E-2],"split_indices":[272,0,0],"split_type":[0,0,0],"sum_hessian":[2.794308E0,1.217783E0,1.576525E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-4.126205E-4,9.989941E-3,-1.2109649E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":436,"left_children":[1,-1,-1],"loss_changes":[6.452384E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.8765771E-1,9.989941E-3,-1.2109649E-2],"split_indices":[224,0,0],"split_type":[0,0,0],"sum_hessian":[2.792554E0,1.6048757E0,1.1876782E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-8.491619E-4,1.0272368E-2,-1.1869042E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":437,"left_children":[1,-1,-1],"loss_changes":[6.504504E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[4.4586396E-1,1.0272368E-2,-1.1869042E-2],"split_indices":[136,0,0],"split_type":[0,0,0],"sum_hessian":[2.7908E0,1.5245234E0,1.2662767E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.0164911E-3,-9.671423E-3,1.2894003E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":438,"left_children":[1,-1,-1],"loss_changes":[6.590356E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.2208195E0,-9.671423E-3,1.2894003E-2],"split_indices":[88,0,0],"split_type":[0,0,0],"sum_hessian":[2.7890291E0,1.787681E0,1.0013481E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.4008835E-4,1.3127896E-2,-9.500945E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":439,"left_children":[1,-1,-1],"loss_changes":[6.6407686E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.8274283E0,1.3127896E-2,-9.500945E-3],"split_indices":[88,0,0],"split_type":[0,0,0],"sum_hessian":[2.787228E0,1.0169992E0,1.7702289E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.0570167E-3,-1.0119039E-2,1.2119727E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":440,"left_children":[1,-1,-1],"loss_changes":[6.5434203E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.8292725E0,-1.0119039E-2,1.2119727E-2],"split_indices":[32,0,0],"split_type":[0,0,0],"sum_hessian":[2.785445E0,1.5540043E0,1.2314407E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.2696116E-3,-1.2540627E-2,1.0068107E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":441,"left_children":[1,-1,-1],"loss_changes":[6.6653E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.236605E-1,-1.2540627E-2,1.0068107E-2],"split_indices":[366,0,0],"split_type":[0,0,0],"sum_hessian":[2.7836797E0,1.0665226E0,1.7171571E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[2.348666E-4,9.856777E-3,-1.2543104E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":442,"left_children":[1,-1,-1],"loss_changes":[6.5607764E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.0193767E-1,9.856777E-3,-1.2543104E-2],"split_indices":[448,0,0],"split_type":[0,0,0],"sum_hessian":[2.78187E0,1.6895661E0,1.0923038E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-5.466213E-4,1.1383549E-2,-1.0538391E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":443,"left_children":[1,-1,-1],"loss_changes":[6.364621E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.2973647E-1,1.1383549E-2,-1.0538391E-2],"split_indices":[371,0,0],"split_type":[0,0,0],"sum_hessian":[2.7801023E0,1.2696303E0,1.5104719E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.5885197E-4,-1.2312052E-2,9.644622E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":444,"left_children":[1,-1,-1],"loss_changes":[6.3097845E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.0950794E-1,-1.2312052E-2,9.644622E-3],"split_indices":[183,0,0],"split_type":[0,0,0],"sum_hessian":[2.778379E0,1.107137E0,1.671242E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-8.189041E-4,-1.22822E-2,9.548459E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":445,"left_children":[1,-1,-1],"loss_changes":[6.2499773E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.0963696E-1,-1.22822E-2,9.548459E-3],"split_indices":[454,0,0],"split_type":[0,0,0],"sum_hessian":[2.7766733E0,1.1317583E0,1.644915E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.2977998E-3,9.576977E-3,-1.2007315E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":446,"left_children":[1,-1,-1],"loss_changes":[6.1343536E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[7.566936E-1,9.576977E-3,-1.2007315E-2],"split_indices":[130,0,0],"split_type":[0,0,0],"sum_hessian":[2.7749717E0,1.5882174E0,1.1867543E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.5537811E-3,1.203942E-2,-1.0935268E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":447,"left_children":[1,-1,-1],"loss_changes":[6.9518336E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-8.1226087E-1,1.203942E-2,-1.0935268E-2],"split_indices":[86,0,0],"split_type":[0,0,0],"sum_hessian":[2.7733126E0,1.195398E0,1.5779146E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-7.949222E-4,1.145537E-2,-1.0178904E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":448,"left_children":[1,-1,-1],"loss_changes":[6.1666453E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.6711326E0,1.145537E-2,-1.0178904E-2],"split_indices":[149,0,0],"split_type":[0,0,0],"sum_hessian":[2.7714045E0,1.2033685E0,1.568036E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.1820405E-4,9.753304E-3,-1.1909415E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":449,"left_children":[1,-1,-1],"loss_changes":[6.1616697E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.0457262E-1,9.753304E-3,-1.1909415E-2],"split_indices":[168,0,0],"split_type":[0,0,0],"sum_hessian":[2.7697377E0,1.6108456E0,1.1588922E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-7.252421E-4,-1.0067016E-2,1.1444267E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":450,"left_children":[1,-1,-1],"loss_changes":[6.08911E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.3680219E-1,-1.0067016E-2,1.1444267E-2],"split_indices":[247,0,0],"split_type":[0,0,0],"sum_hessian":[2.7680688E0,1.5747827E0,1.1932861E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.4239407E-4,-1.1157573E-2,1.00807855E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":451,"left_children":[1,-1,-1],"loss_changes":[5.9586465E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-8.7341E-2,-1.1157573E-2,1.00807855E-2],"split_indices":[298,0,0],"split_type":[0,0,0],"sum_hessian":[2.7664084E0,1.2699511E0,1.4964573E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-3.8591126E-4,1.2249544E-2,-9.314917E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":452,"left_children":[1,-1,-1],"loss_changes":[6.026105E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.64045E-1,1.2249544E-2,-9.314917E-3],"split_indices":[390,0,0],"split_type":[0,0,0],"sum_hessian":[2.7647889E0,1.0379713E0,1.7268176E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[5.1065406E-4,-9.389927E-3,1.1836751E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":453,"left_children":[1,-1,-1],"loss_changes":[5.8969962E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.019842E0,-9.389927E-3,1.1836751E-2],"split_indices":[150,0,0],"split_type":[0,0,0],"sum_hessian":[2.7631607E0,1.6289486E0,1.1342121E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.0085442E-3,1.4233152E-2,-1.4350493E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":454,"left_children":[1,-1,-1],"loss_changes":[1.080093E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.3998775E-1,1.4233152E-2,-1.4350493E-2],"split_indices":[159,0,0],"split_type":[0,0,0],"sum_hessian":[2.761559E0,1.4303699E0,1.3311892E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[6.757092E-4,-1.1561352E-2,1.0350389E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":455,"left_children":[1,-1,-1],"loss_changes":[6.3150874E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.5198672E-1,-1.1561352E-2,1.0350389E-2],"split_indices":[495,0,0],"split_type":[0,0,0],"sum_hessian":[2.758597E0,1.2130333E0,1.5455636E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.5079851E-4,-1.1072336E-2,1.0193534E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":456,"left_children":[1,-1,-1],"loss_changes":[5.9637357E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-4.559751E-2,-1.1072336E-2,1.0193534E-2],"split_indices":[182,0,0],"split_type":[0,0,0],"sum_hessian":[2.7568984E0,1.2721689E0,1.4847295E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.2900705E-4,-1.0380722E-2,1.0996865E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":457,"left_children":[1,-1,-1],"loss_changes":[6.0304943E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.0041963E0,-1.0380722E-2,1.0996865E-2],"split_indices":[78,0,0],"split_type":[0,0,0],"sum_hessian":[2.75527E0,1.4529616E0,1.3023084E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[7.372103E-5,-1.150433E-2,9.72743E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":458,"left_children":[1,-1,-1],"loss_changes":[5.909064E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.8817723E0,-1.150433E-2,9.72743E-3],"split_indices":[76,0,0],"split_type":[0,0,0],"sum_hessian":[2.7536123E0,1.1739795E0,1.5796328E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-4.2210147E-4,9.890692E-3,-1.1239319E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":459,"left_children":[1,-1,-1],"loss_changes":[5.8759367E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.5563698E-1,9.890692E-3,-1.1239319E-2],"split_indices":[485,0,0],"split_type":[0,0,0],"sum_hessian":[2.7519946E0,1.5051603E0,1.2468343E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-6.551449E-4,-1.0280101E-2,1.1014396E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":460,"left_children":[1,-1,-1],"loss_changes":[5.9688776E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.3398013E-1,-1.0280101E-2,1.1014396E-2],"split_indices":[380,0,0],"split_type":[0,0,0],"sum_hessian":[2.750414E0,1.4917265E0,1.2586874E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.629185E-4,1.1265339E-2,-1.0203116E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":461,"left_children":[1,-1,-1],"loss_changes":[6.061074E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.276636E-1,1.1265339E-2,-1.0203116E-2],"split_indices":[329,0,0],"split_type":[0,0,0],"sum_hessian":[2.7487965E0,1.2431437E0,1.5056528E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[9.9423676E-5,-1.1581543E-2,9.755221E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":462,"left_children":[1,-1,-1],"loss_changes":[5.9569846E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.371829E-1,-1.1581543E-2,9.755221E-3],"split_indices":[230,0,0],"split_type":[0,0,0],"sum_hessian":[2.7471561E0,1.1651732E0,1.581983E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-4.183559E-4,9.593459E-3,-1.15401335E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":463,"left_children":[1,-1,-1],"loss_changes":[5.847051E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.9867137E-1,9.593459E-3,-1.15401335E-2],"split_indices":[179,0,0],"split_type":[0,0,0],"sum_hessian":[2.745547E0,1.5690925E0,1.1764545E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-8.1351324E-4,-1.1349233E-2,9.753272E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":464,"left_children":[1,-1,-1],"loss_changes":[5.8483877E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.6846621E-1,-1.1349233E-2,9.753272E-3],"split_indices":[362,0,0],"split_type":[0,0,0],"sum_hessian":[2.7439647E0,1.2358913E0,1.5080733E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-9.9664E-4,1.5041514E-2,-2.0532653E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":465,"left_children":[1,-1,-1],"loss_changes":[1.6338388E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.0011115E0,1.5041514E-2,-2.0532653E-2],"split_indices":[117,0,0],"split_type":[0,0,0],"sum_hessian":[2.7423744E0,1.7057439E0,1.0366305E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.1708275E-3,-1.2386446E-2,9.056785E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":466,"left_children":[1,-1,-1],"loss_changes":[5.9780595E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.6732554E0,-1.2386446E-2,9.056785E-3],"split_indices":[71,0,0],"split_type":[0,0,0],"sum_hessian":[2.7379107E0,1.1146321E0,1.6232786E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.412314E-3,1.1611872E-2,-9.860289E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":467,"left_children":[1,-1,-1],"loss_changes":[5.951038E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.2231853E0,1.1611872E-2,-9.860289E-3],"split_indices":[25,0,0],"split_type":[0,0,0],"sum_hessian":[2.7362742E0,1.0490303E0,1.6872439E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.2201433E-3,1.1265985E-2,-1.0014126E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":468,"left_children":[1,-1,-1],"loss_changes":[5.910515E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.111781E-1,1.1265985E-2,-1.0014126E-2],"split_indices":[326,0,0],"split_type":[0,0,0],"sum_hessian":[2.7346475E0,1.1638193E0,1.5708282E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-5.189458E-4,-9.484942E-3,1.1732604E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":469,"left_children":[1,-1,-1],"loss_changes":[5.8367946E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.1811806E-1,-9.484942E-3,1.1732604E-2],"split_indices":[439,0,0],"split_type":[0,0,0],"sum_hessian":[2.73305E0,1.6446121E0,1.088438E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[2.3953743E-4,-1.0998493E-2,1.0193509E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":470,"left_children":[1,-1,-1],"loss_changes":[5.8913864E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.0751958E-2,-1.0998493E-2,1.0193509E-2],"split_indices":[357,0,0],"split_type":[0,0,0],"sum_hessian":[2.7314656E0,1.2632165E0,1.4682491E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-4.8333233E-5,9.569084E-3,-1.1524428E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":471,"left_children":[1,-1,-1],"loss_changes":[5.7967254E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.796376E-1,9.569084E-3,-1.1524428E-2],"split_indices":[233,0,0],"split_type":[0,0,0],"sum_hessian":[2.7298582E0,1.5815912E0,1.1482669E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-5.587441E-4,-1.1744943E-2,9.3348725E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":472,"left_children":[1,-1,-1],"loss_changes":[5.775557E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.0657914E0,-1.1744943E-2,9.3348725E-3],"split_indices":[94,0,0],"split_type":[0,0,0],"sum_hessian":[2.7282963E0,1.1235001E0,1.6047962E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.0382405E-3,-1.4153949E-2,1.864895E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":473,"left_children":[1,-1,-1],"loss_changes":[1.3800742E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.123695E0,-1.4153949E-2,1.864895E-2],"split_indices":[156,0,0],"split_type":[0,0,0],"sum_hessian":[2.7267146E0,1.7225958E0,1.0041188E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[4.820773E-4,1.1454364E-2,-9.827439E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":474,"left_children":[1,-1,-1],"loss_changes":[5.9160665E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.058481E0,1.1454364E-2,-9.827439E-3],"split_indices":[29,0,0],"split_type":[0,0,0],"sum_hessian":[2.7229445E0,1.2062451E0,1.5166994E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[7.749452E-4,-1.1560449E-2,9.765532E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":475,"left_children":[1,-1,-1],"loss_changes":[5.902865E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.5414672E-1,-1.1560449E-2,9.765532E-3],"split_indices":[218,0,0],"split_type":[0,0,0],"sum_hessian":[2.72135E0,1.121419E0,1.599931E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[5.9575847E-5,-1.1167935E-2,9.941992E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":476,"left_children":[1,-1,-1],"loss_changes":[5.82176E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.0569758E-1,-1.1167935E-2,9.941992E-3],"split_indices":[462,0,0],"split_type":[0,0,0],"sum_hessian":[2.7197504E0,1.2196779E0,1.5000725E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.8816133E-4,9.147641E-3,-1.1975898E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":477,"left_children":[1,-1,-1],"loss_changes":[5.7529206E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[3.1681415E-1,9.147641E-3,-1.1975898E-2],"split_indices":[333,0,0],"split_type":[0,0,0],"sum_hessian":[2.7181723E0,1.6597306E0,1.0584418E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-9.63364E-4,1.0030477E-2,-1.0762657E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":478,"left_children":[1,-1,-1],"loss_changes":[5.66287E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[9.095875E-2,1.0030477E-2,-1.0762657E-2],"split_indices":[189,0,0],"split_type":[0,0,0],"sum_hessian":[2.7166343E0,1.389701E0,1.3269333E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-8.7144756E-4,1.2618448E-2,-1.1047292E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":479,"left_children":[1,-1,-1],"loss_changes":[7.283369E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.7596493E0,1.2618448E-2,-1.1047292E-2],"split_indices":[22,0,0],"split_type":[0,0,0],"sum_hessian":[2.7150989E0,1.1599925E0,1.5551064E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.8737366E-4,1.183696E-2,-9.167899E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":480,"left_children":[1,-1,-1],"loss_changes":[5.676657E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.7445644E-1,1.183696E-2,-9.167899E-3],"split_indices":[490,0,0],"split_type":[0,0,0],"sum_hessian":[2.7131195E0,1.0471779E0,1.6659416E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[5.8852806E-4,-1.1745999E-2,1.1170702E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":481,"left_children":[1,-1,-1],"loss_changes":[6.8635377E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.83142E0,-1.1745999E-2,1.1170702E-2],"split_indices":[31,0,0],"split_type":[0,0,0],"sum_hessian":[2.711578E0,1.2680544E0,1.4435235E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[2.5672402E-4,-9.097402E-3,1.1888509E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":482,"left_children":[1,-1,-1],"loss_changes":[5.6683584E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.5862882E-1,-9.097402E-3,1.1888509E-2],"split_indices":[346,0,0],"split_type":[0,0,0],"sum_hessian":[2.7097125E0,1.6544354E0,1.0552771E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[9.3328E-4,1.1189017E-2,-9.465675E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":483,"left_children":[1,-1,-1],"loss_changes":[5.5572097E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.9723969E-1,1.1189017E-2,-9.465675E-3],"split_indices":[95,0,0],"split_type":[0,0,0],"sum_hessian":[2.7081754E0,1.2079384E0,1.500237E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.1189235E-3,-1.15597E-2,9.300801E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":484,"left_children":[1,-1,-1],"loss_changes":[5.586735E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.7680993E-1,-1.15597E-2,9.300801E-3],"split_indices":[268,0,0],"split_type":[0,0,0],"sum_hessian":[2.706664E0,1.0388534E0,1.6678107E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.7339361E-4,9.725461E-3,-1.0976022E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":485,"left_children":[1,-1,-1],"loss_changes":[5.5778855E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[7.9127884E-1,9.725461E-3,-1.0976022E-2],"split_indices":[37,0,0],"split_type":[0,0,0],"sum_hessian":[2.705157E0,1.5040061E0,1.2011509E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.1652553E-4,-9.848154E-3,1.1013856E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":486,"left_children":[1,-1,-1],"loss_changes":[5.6654373E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[4.413488E-1,-9.848154E-3,1.1013856E-2],"split_indices":[158,0,0],"split_type":[0,0,0],"sum_hessian":[2.7036376E0,1.4947628E0,1.2088748E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.6297247E-4,1.183392E-2,-8.962025E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":487,"left_children":[1,-1,-1],"loss_changes":[5.5466206E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.3680513E-1,1.183392E-2,-8.962025E-3],"split_indices":[288,0,0],"split_type":[0,0,0],"sum_hessian":[2.7020955E0,1.0350746E0,1.6670209E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[8.9132093E-4,1.036491E-2,-1.0235555E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":488,"left_children":[1,-1,-1],"loss_changes":[5.539468E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.1725941E-2,1.036491E-2,-1.0235555E-2],"split_indices":[491,0,0],"split_type":[0,0,0],"sum_hessian":[2.7005937E0,1.3835729E0,1.3170208E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[6.595043E-4,-1.0865286E-2,9.80116E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":489,"left_children":[1,-1,-1],"loss_changes":[5.54994E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.8090875E-1,-1.0865286E-2,9.80116E-3],"split_indices":[413,0,0],"split_type":[0,0,0],"sum_hessian":[2.6990972E0,1.1931556E0,1.5059415E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.7349565E-4,-1.1521515E-2,9.242723E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":490,"left_children":[1,-1,-1],"loss_changes":[5.553328E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.2397411E-1,-1.1521515E-2,9.242723E-3],"split_indices":[436,0,0],"split_type":[0,0,0],"sum_hessian":[2.6975992E0,1.0817596E0,1.6158396E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-4.954092E-4,1.1682773E-2,-9.028869E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":491,"left_children":[1,-1,-1],"loss_changes":[5.48683E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.408786E0,1.1682773E-2,-9.028869E-3],"split_indices":[80,0,0],"split_type":[0,0,0],"sum_hessian":[2.696092E0,1.0206546E0,1.6754373E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[3.71994E-4,1.1645369E-2,-9.01454E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":492,"left_children":[1,-1,-1],"loss_changes":[5.487406E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-2.2420745E-1,1.1645369E-2,-9.01454E-3],"split_indices":[180,0,0],"split_type":[0,0,0],"sum_hessian":[2.6945984E0,1.0683516E0,1.6262468E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[9.734139E-4,-1.0034097E-2,1.0466856E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":493,"left_children":[1,-1,-1],"loss_changes":[5.4783146E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[5.3132057E-2,-1.0034097E-2,1.0466856E-2],"split_indices":[49,0,0],"split_type":[0,0,0],"sum_hessian":[2.693115E0,1.3434852E0,1.3496298E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[7.989865E-4,1.2977481E-2,-1.5937798E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":494,"left_children":[1,-1,-1],"loss_changes":[1.0750559E-2,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.7981895E-1,1.2977481E-2,-1.5937798E-2],"split_indices":[310,0,0],"split_type":[0,0,0],"sum_hessian":[2.6916325E0,1.6165807E0,1.0750518E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.4401594E-4,8.782915E-3,-1.1870624E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":495,"left_children":[1,-1,-1],"loss_changes":[5.440606E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.7524998E0,8.782915E-3,-1.1870624E-2],"split_indices":[132,0,0],"split_type":[0,0,0],"sum_hessian":[2.6887276E0,1.6817724E0,1.0069553E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-1.00958E-3,1.1127018E-2,-9.397944E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":496,"left_children":[1,-1,-1],"loss_changes":[5.4208697E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-3.244198E-1,1.1127018E-2,-9.397944E-3],"split_indices":[240,0,0],"split_type":[0,0,0],"sum_hessian":[2.6872184E0,1.0917678E0,1.5954506E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-2.3715824E-4,9.525607E-3,-1.0930805E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":497,"left_children":[1,-1,-1],"loss_changes":[5.4249116E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[1.4130208E0,9.525607E-3,-1.0930805E-2],"split_indices":[120,0,0],"split_type":[0,0,0],"sum_hessian":[2.6857495E0,1.4909931E0,1.1947564E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[-5.474399E-4,-9.304164E-3,1.1309201E-2],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[true,false,false],"id":498,"left_children":[1,-1,-1],"loss_changes":[5.4620067E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[2.356022E-1,-9.304164E-3,1.1309201E-2],"split_indices":[211,0,0],"split_type":[0,0,0],"sum_hessian":[2.6842878E0,1.599315E0,1.0849727E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}},{"base_weights":[1.6103139E-4,-1.138246E-2,9.320275E-3],"categories":[],"categories_nodes":[],"categories_segments":[],"categories_sizes":[],"default_left":[false,false,false],"id":499,"left_children":[1,-1,-1],"loss_changes":[5.515701E-3,0E0,0E0],"parents":[2147483647,0,0],"right_children":[2,-1,-1],"split_conditions":[-1.7813606E0,-1.138246E-2,9.320275E-3],"split_indices":[10,0,0],"split_type":[0,0,0],"sum_hessian":[2.6828132E0,1.099587E0,1.5832262E0],"tree_param":{"num_deleted":"0","num_feature":"500","num_nodes":"3","size_leaf_vector":"0"}}]},"name":"gbtree"},"learner_model_param":{"base_score":"5E-1","num_class":"0","num_feature":"500"},"objective":{"name":"binary:logistic","reg_loss_param":{"scale_pos_weight":"1"}}},"version":[1,3,3]} | 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart/.helmignore | # Patterns to ignore when building packages.
# This supports shell glob matching, relative path matching, and
# negation (prefixed with !). Only one pattern per line.
.DS_Store
# Common VCS dirs
.git/
.gitignore
.bzr/
.bzrignore
.hg/
.hgignore
.svn/
# Common backup files
*.swp
*.bak
*.tmp
*~
# Various IDEs
.project
.idea/
*.tmproj
.vscode/
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart/triton/Chart.yaml | # Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
apiVersion: v1
appVersion: "2.8"
description: Triton Inference Server
name: triton-inference-server
version: 2.8.7
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart/triton/values.yaml | # Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
initReplicaCount: 1
minReplicaCount: 1
maxReplicaCount: 3
HPATargetAverageValue: 85
modelRepositoryPath: ''
publishedVersion: '2.8.7'
image:
registry: gcr.io
repository: ''
tag: 'latest'
pullPolicy: IfNotPresent
# modify the model repository here to match your GCP storage bucket
numGpus: 1
strictModelConfig: False
# add in custom library which could include custom ops in the model
ldPreloadPath: ''
logVerboseLevel: 0
allowGPUMetrics: True
service:
type: NodePort
deployment:
livenessProbe:
failureThreshold: 60
initialDelaySeconds: 10
periodSeconds: 5
successThreshold: 1
timeoutSeconds: 1
readinessProbe:
failureThreshold: 60
initialDelaySeconds: 10
periodSeconds: 5
successThreshold: 1
timeoutSeconds: 1
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart/triton | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart/triton/templates/_helpers.tpl | {{/*
# Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/}}
{{/* vim: set filetype=mustache: */}}
{{/*
Expand the name of the chart.
*/}}
{{- define "triton-inference-server.name" -}}
{{- default .Chart.Name .Values.nameOverride | trunc 63 | trimSuffix "-" -}}
{{- end -}}
{{/*
Create a default fully qualified app name.
We truncate at 63 chars because some Kubernetes name fields are limited to this (by the DNS naming spec).
If release name contains chart name it will be used as a full name.
*/}}
{{- define "triton-inference-server.fullname" -}}
{{- if .Values.fullnameOverride -}}
{{- .Values.fullnameOverride | trunc 63 | trimSuffix "-" -}}
{{- else -}}
{{- $name := default .Chart.Name .Values.nameOverride -}}
{{- if contains $name .Release.Name -}}
{{- .Release.Name | trunc 63 | trimSuffix "-" -}}
{{- else -}}
{{- printf "%s-%s" .Release.Name $name | trunc 63 | trimSuffix "-" -}}
{{- end -}}
{{- end -}}
{{- end -}}
{{/*
Create chart name and version as used by the chart label.
*/}}
{{- define "triton-inference-server.chart" -}}
{{- printf "%s-%s" .Chart.Name .Chart.Version | replace "+" "_" | trunc 63 | trimSuffix "-" -}}
{{- end -}}
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart/triton | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart/triton/templates/service.yaml | # Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
apiVersion: v1
kind: Service
metadata:
name: {{ template "triton-inference-server.name" . }}
namespace: {{ .Release.Namespace }}
annotations:
cloud.google.com/neg: '{"ingress": true}'
labels:
app: {{ template "triton-inference-server.name" . }}
chart: {{ template "triton-inference-server.chart" . }}
release: {{ .Release.Name }}
heritage: {{ .Release.Service }}
spec:
type: {{ .Values.service.type }}
ports:
- port: 8000
targetPort: http
name: http-inference-server
- port: 8001
targetPort: grpc
name: grpc-inference-server
- port: 8002
targetPort: metrics
name: metrics-inference-server
selector:
app: {{ template "triton-inference-server.name" . }}
release: {{ .Release.Name }}
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart/triton | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart/triton/templates/hpa.yaml | # Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
apiVersion: autoscaling/v2beta1
kind: HorizontalPodAutoscaler
metadata:
name: triton-hpa
namespace: {{ .Release.Namespace }}
labels:
app: triton-hpa
spec:
minReplicas: {{ .Values.minReplicaCount }}
maxReplicas: {{ .Values.maxReplicaCount }}
metrics:
- type: External
external:
metricName: kubernetes.io|container|accelerator|duty_cycle
targetAverageValue: {{ .Values.HPATargetAverageValue }}
scaleTargetRef:
apiVersion: apps/v1
kind: Deployment
name: {{ template "triton-inference-server.name" . }}
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart/triton | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart/triton/templates/istio-gateway.yaml | # Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
apiVersion: networking.istio.io/v1alpha3
kind: Gateway
metadata:
name: triton-gateway
spec:
selector:
istio: ingressgateway # use istio default controller
servers:
- port:
number: 80
name: http
protocol: HTTP
hosts:
- "*"
- port:
number: 31400
name: grpc
protocol: TCP
hosts:
- "*"
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart/triton | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart/triton/templates/deployment.yaml | # Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
apiVersion: apps/v1
kind: Deployment
metadata:
name: {{ template "triton-inference-server.name" . }}
namespace: {{ .Release.Namespace }}
labels:
app: {{ template "triton-inference-server.name" . }}
chart: {{ template "triton-inference-server.chart" . }}
release: {{ .Release.Name }}
heritage: {{ .Release.Service }}
spec:
replicas: {{ .Values.initReplicaCount }}
selector:
matchLabels:
app: {{ template "triton-inference-server.name" . }}
release: {{ .Release.Name }}
template:
metadata:
labels:
app: {{ template "triton-inference-server.name" . }}
release: {{ .Release.Name }}
spec:
volumes:
- name: gcsfs-creds
secret:
secretName: gcsfs-creds
items:
- key: keyfile.json
path: keyfile.json
containers:
- name: {{ .Chart.Name }}
image: "{{ .Values.image.registry }}/{{ .Values.image.repository }}:{{ .Values.image.tag }}"
imagePullPolicy: {{ .Values.image.pullPolicy }}
resources:
limits:
nvidia.com/gpu: {{ .Values.image.numGpus }}
volumeMounts:
- name: gcsfs-creds
mountPath: "/etc/secrets"
readOnly: true
env:
- name: LD_PRELOAD
value: {{ .Values.image.ldPreloadPath }}
- name: GOOGLE_APPLICATION_CREDENTIALS
value: "/etc/secrets/keyfile.json"
args: ["tritonserver", "--model-store={{ .Values.modelRepositoryPath }}",
"--strict-model-config={{ .Values.image.strictModelConfig }}",
"--log-verbose={{ .Values.image.logVerboseLevel }}",
"--allow-gpu-metrics={{ .Values.image.allowGPUMetrics }}"]
ports:
- containerPort: 8000
name: http
- containerPort: 8001
name: grpc
- containerPort: 8002
name: metrics
livenessProbe:
httpGet:
path: /v2/health/live
port: http
initialDelaySeconds: {{ .Values.deployment.livenessProbe.initialDelaySeconds }}
periodSeconds: {{ .Values.deployment.livenessProbe.periodSeconds }}
timeoutSeconds: {{ .Values.deployment.livenessProbe.timeoutSeconds }}
successThreshold: {{ .Values.deployment.livenessProbe.successThreshold }}
failureThreshold: {{ .Values.deployment.livenessProbe.failureThreshold }}
readinessProbe:
httpGet:
path: /v2/health/ready
port: http
initialDelaySeconds: {{ .Values.deployment.readinessProbe.initialDelaySeconds }}
periodSeconds: {{ .Values.deployment.readinessProbe.periodSeconds }}
timeoutSeconds: {{ .Values.deployment.readinessProbe.timeoutSeconds }}
successThreshold: {{ .Values.deployment.readinessProbe.successThreshold }}
failureThreshold: {{ .Values.deployment.readinessProbe.failureThreshold }}
securityContext:
runAsUser: 1000
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart/triton | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/GCP/FIL/helm/chart/triton/templates/istio-vs.yaml | # Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: triton-vs
spec:
hosts:
- "*"
gateways:
- triton-gateway
tcp:
- match:
- port: 31400
route:
- destination:
host: {{ template "triton-inference-server.name" . }}
port:
number: 8001
http:
- match:
- port: 80
route:
- destination:
host: {{ template "triton-inference-server.name" . }}
port:
number: 8000
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/README.md | ## Deploying a Triton service that uses a custom plugin for the RAPIDS cuML Forest Inference Library (FIL).
**NOTE:** For steps to setup a horizontally autoscalable Triton Service in EKS refer to the instructions in [Detailed_HPA_Setup](./Detailed_HPA_Setup.md).
### Overview
This example will illustrate the workflow to deploy a [Triton Inference Server](https://developer.nvidia.com/nvidia-triton-inference-server), with the [cuML Forest Inference Library (FIL) backend](https://github.com/triton-inference-server/fil_backend) on [Amazon Elastic Kubernetes Service (EKS)](https://aws.amazon.com/eks). [FIL](https://docs.rapids.ai/api/cuml/stable/api.html?highlight=forestinference#cuml.ForestInference) allows GPU accelerated inference for random forest and boosted decision tree models, and the FIL backend is right now available as a fully integrated part of Triton.
It is assumed that you have an existing account with sufficient compute and GPU quota, a correctly configured `kubectl` utility, and a properly configured and accessible AWS S3 bucket.
**NOTE:** When referring to configuration parameters or elements that are `specific to your environment`, they will be represented as linux-style environment variables ex. ${YOUR_CONFIG_PARAM}.
Specific parameters used here include:
- REGION_NAME= < your preferred location >
- EKS_CLUSTER_NAME= < your cluster name >
- INSTANCE_TYPE= g4dn.12xlarge # or any other VM size with GPUs.
- AWS_ACCOUNT_ID= < select your ACCOUNT ID >
- REPOSITORY_NAME= titon_fil_backend # < name of the custom image we are going to build for triton backend, you can change this to your preference >
- S3_BUCKET_PATH= < s3 bucket name (existing or to be created) to host the models >
---
### Step 1: Prerequisites
- AWS account with at least the following permissions mentioned in https://eksctl.io/usage/minimum-iam-policies/ along with permissions to create, pull and push to AWS ECR repositories.
- System Softwares
- `docker`
- `aws-cli` , `eksctl`
- `helm` , `kubectl` , `istioctl`
- Python Libraries
```shell
pip install nvidia-pyindex
pip install tritonclient[all]
```
- Login credentials for aws-cli and kubectl configured.
**Note:** The rest of the demo uses `aws-cli` [version 2](https://docs.aws.amazon.com/cli/latest/userguide/install-cliv2.html). Hence some commands will be a little different from `aws-cli` version 1. We will also be using `helm` 3 that does not need `tiller`.
### Step 2: Obtain the Triton Inference Server official NVIDIA image (and optionally build a custom one)
The [FIL backend plugin](https://github.com/triton-inference-server/fil_backend) is natively supported in the official [Triton Inference Server](https://ngc.nvidia.com/catalog/containers/nvidia:tritonserver) image available in the [NVIDIA NGC Catalog](https://ngc.nvidia.com/catalog). The images in this catalogue are the preferred way of deploying the FIL backend. For this example, we will use the most current image: `nvcr.io/nvidia/tritonserver:21.06.1-py3`.
**You should skip to [**Step 3**](#Step-3:-Create-some-Triton-model-repository-entries,-or-use-the-provided-examples) if you are using the official NVIDIA image.**
For informational purpose, we will also show how to create a custom image with FIL backend below.
#### Step 2.a: Create an ECR repository
```shell
aws ecr get-login-password --region ${REGION_NAME} | docker login --username AWS --password-stdin ${AWS_ACCOUNT_ID}.dkr.ecr.${REGION_NAME}.amazonaws.com
aws ecr create-repository \
--repository-name ${REPOSITORY_NAME} \
--image-scanning-configuration scanOnPush=true \
--region ${REGION_NAME}
```
#### Step 2.b: Create and push the image to the ECR repository
**On a different terminal** do the following:
```shell
git clone https://github.com/triton-inference-server/fil_backend.git
cd fil_backend
docker build --tag ${AWS_ACCOUNT_ID}.dkr.ecr.${REGION_NAME}.amazonaws.com/${REPOSITORY_NAME} -f ops/Dockerfile .
docker push ${AWS_ACCOUNT_ID}.dkr.ecr.${REGION_NAME}.amazonaws.com/${REPOSITORY_NAME}:latest
```
### Step 3: Create some Triton model repository entries, or use the provided examples
A set of sample models, along with their `.pbtext` definition are provided in `./model_repository`. The layout structure and requirements are defined in the [Triton server docs](https://github.com/triton-inference-server/server/blob/master/docs/model_configuration.md),
and a brief introduction is also provided with the [FIL backend implementation](https://github.com/triton-inference-server/fil_backend/blob/main/README.md).
For the purpose of this demo we will use the models in `.model_repository` directory; however, feel free to add additional models. Make sure they follow the same structure in the docs. Here we assume the AWS `S3_BUCKET_PATH` bucket exists. You can create one in case it does not exist.
```shell
aws s3 cp --recursive ./model_repository s3://${S3_BUCKET_PATH}/model_repository/
aws s3 ls s3://${S3_BUCKET_PATH}/model_repository/
```
### Step 4: Configure an EKS cluster
We first need to create a Kubernetes cluster that will host our
Triton service, and has a GPU node pool with enough nodes and GPUs to illustrate horizontal scaling.
Lets first create a minimal yaml configuration file with name `eksctl_config.yaml` with the following in it
```yaml
apiVersion: eksctl.io/v1alpha5
kind: ClusterConfig
metadata:
name: <name set in EKS_CLUSTER_NAME>
region: <region set in REGION_NAME>
version: "1.20"
managedNodeGroups:
- name: nodegroup-managed-gpu-dev
minSize: 0
desiredCapacity: 2
maxSize: 4
instanceType: <instance type set in INSTANCE_TYPE>
ssh:
allow: true
```
To make things simple, we will use a [managed node group](https://docs.aws.amazon.com/eks/latest/userguide/managed-node-groups.html) to let EKS perform the provisioning and lifecycle management on our behalf. We are also allowing `ssh` access into the nodes for debugging. By default this will use the `~/.ssh/id_rsa.pub` file in your local machine.
With the `yaml` file, creating the EKS cluster is as simple as running the following command:
```shell
eksctl create cluster -f eksctl_config.yaml
```
This will take a few minutes before it completes. Grab a coffee :coffee: :coffee: . For more `yaml` configurations, you can refer to [eksctl documentation](https://eksctl.io/usage/schema/).
You can check whether the cluster is successfully created with following:
```shell
eksctl get cluster --name $EKS_CLUSTER_NAME --region $REGION_NAME
```
**NOTE:** For more `eksctl` related references on how to delete or modify the cluster, visit https://docs.aws.amazon.com/eks/latest/userguide/getting-started-eksctl.html.
Once the cluster is created successfully, let's get the credentials for your EKS cluster to access it from your machine.
```shell
aws eks update-kubeconfig --name $EKS_CLUSTER_NAME --region $REGION_NAME
```
Check whether you are able to access the nodes:
```shell
>> kubectl get nodes
NAME STATUS ROLES AGE VERSION
ip-172-31-12-100.ec2.internal NotReady <none> 10m v1.20.4-eks-6b7464
ip-172-31-34-168.ec2.internal NotReady <none> 10m v1.20.4-eks-6b7464
```
### Step 5: Setting up the EKS cluster to use GPUs for our workload
The good thing about using `eksctl` is that we can simply use a GPU compatible VM instance type with EKS and the AWS AMI resolvers will automatically select the correct EKS optimized accelerated AMI instance. Subsequently, `eksctl` will install the NVIDIA Kubernetes device plugin automatically ([reference](https://eksctl.io/usage/gpu-support/)) in the VMs. Therefore we do not have to do anything additional in this step.
### Step 6: To access models from AWS S3 and to fetch fil_backend image from ECR repository
To fetch the images from the ECR repository and to load the models from the AWS S3 we will use a config file and a secret. You need to convert the AWS credentials of your account in the base64 format and add it to the `./helm/charts/triton/values.yaml`, [similar to these instructions](https://github.com/triton-inference-server/server/tree/main/deploy/aws).
```shell
echo -n 'AWS_REGION' | base64
echo -n 'AWS_SECRET_KEY_ID' | base64
echo -n 'AWS_SECRET_ACCESS_KEY' | base64
echo -n 'AWS_SESSION_TOKEN' | base64
```
where `AWS_REGION`, `AWS_SECRET_KEY_ID`, `AWS_SECRET_ACCESS_KEY` and `AWS_SESSION_TOKEN` can be obtained either from the environment variable or the aws credentials file. `helm` will populate the `helm/charts/triton/templates/secrets.yaml` appropriately during deployment.
Replace the following fields in the `helm/charts/triton/values.yml` file with the above details
```yaml
image:
imageName: nvcr.io/nvidia/tritonserver:21.06.1-py3 # < update this if you are using a custom image >
pullPolicy: IfNotPresent
modelRepositoryPath: s3://S3_BUCKET_PATH/model_repository
numGpus: 1
logVerboseLevel: 0
allowGPUMetrics: True
secret:
region: < replace with base64 AWS_REGION >
id: < replace with base64 AWS_SECRET_KEY_ID >
key: < replace with base64 AWS_SECRET_ACCESS_KEY >
session_token: < replace with base64 AWS_SESSION_TOKEN >
```
### Step 7: Install `istio` on your EKS cluster
Install the demo profile of `istio` by doing
```shell
istioctl install --set profile=demo
```
This will install the "Istio core", "Istiod", "Ingress gateways" and "Egress gateways" along with exposing all the necessary ports for gRPC and http connections for the ingress gateways.
Finally, enable sidecar injection to the triton pods in the default namespace:
```shell
kubectl label namespace default istio-injection=enabled
```
### Step 8: Install Triton service using helm.
```shell
helm install triton ./helm/charts/triton/
```
This will install a release named `triton`.
When finished, you can delete the triton deployment using
```shell
helm uninstall triton
```
### Step 9: Exploring inference
At this point, your triton inference cluster should be up and running or in process of coming up. Now we can submit some
test data to our running server. The process for doing this, assuming the default model, is illustrated in the jupyter
notebook [triton_inference.ipynb](./triton_inference.ipynb). | 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/triton_inference.ipynb | import os
import numpy
import subprocess
import sys
import time
import tritonclient.http as triton_http
import tritonclient.grpc as triton_grpchttp_port_cmd = "kubectl -n istio-system get service istio-ingressgateway -o jsonpath='{.spec.ports[?(@.name=='http2')].port}'"
grpc_port_cmd = "kubectl -n istio-system get service istio-ingressgateway -o jsonpath='{.spec.ports[?(@.name=='tcp')].port}'"
host_cmd = "kubectl -n istio-system get service istio-ingressgateway -o jsonpath='{.status.loadBalancer.ingress[0].hostname}'"
http_port = subprocess.check_output(http_port_cmd.split()).decode('utf-8').replace("'", "")
grpc_port = subprocess.check_output(grpc_port_cmd.split()).decode('utf-8').replace("'", "")
host = subprocess.check_output(host_cmd.split()).decode('utf-8').replace("'", "")
print(host, http_port)
print(host, grpc_port)
# Set up both HTTP and GRPC clients. Note that the GRPC client is generally
# somewhat faster.
# Generate dummy data to classify
features = 32
samples = 8_000
data = numpy.random.rand(samples, features).astype('float32')http_client = triton_http.InferenceServerClient(
url=f'{host}:{http_port}',
verbose=False,
concurrency=12
)
while (not (http_client.is_server_ready() or http_client.is_model_ready('xgb_model'))):
print("Waiting on server ready")
time.sleep(5)
print(f"Is Server Ready: {http_client.is_server_ready()}")
print(f"Is FIL model ready: {http_client.is_model_ready('xgb_model')}")%%time
# Set up Triton input and output objects for both HTTP and GRPC
triton_input_http = triton_http.InferInput(
'input__0',
(samples, features),
'FP32'
)
triton_input_http.set_data_from_numpy(data, binary_data=True)
triton_output_http = triton_http.InferRequestedOutput(
'output__0',
binary_data=True
)
# Submit inference requests (both HTTP and GRPC)
request_http = http_client.infer(
'xgb_model',
model_version='1',
inputs=[triton_input_http],
outputs=[triton_output_http]
)result_http = request_http.as_numpy('output__0')
result_httpgrpc_client = triton_grpc.InferenceServerClient(
url=f'{host}:{grpc_port}',
verbose = False
)
while (not (grpc_client.is_server_ready() or grpc_client.is_model_ready('xgb_model'))):
print("Waiting on server ready")
time.sleep(5)
print(f"Is Server Ready: {grpc_client.is_server_ready()}")
print(f"Is FIL model ready: {grpc_client.is_model_ready('xgb_model')}")%%time
triton_input_grpc = triton_grpc.InferInput(
'input__0',
[samples, features],
'FP32'
)
triton_input_grpc.set_data_from_numpy(data)
triton_output_grpc = triton_grpc.InferRequestedOutput('output__0')
request_grpc = grpc_client.infer(
'xgb_model',
model_version='1',
inputs=[triton_input_grpc],
outputs=[triton_output_grpc]
)result_grpc = request_grpc.as_numpy('output__0')
result_grpc# Check that we got the same result with both GRPC and HTTP
numpy.testing.assert_almost_equal(result_http, result_grpc)http_client.get_inference_statistics('xgb_model')http_client.get_model_repository_index() | 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/Detailed_HPA_Setup.md |
## Deploying an autoscaling Triton service that uses a custom plugin for the RAPIDS cuML Forest Inference Library (FIL).
### Overview
This example will illustrate the workflow to deploy a [Triton Inference Server](https://developer.nvidia.com/nvidia-triton-inference-server), with the [cuML Forest Inference Library (FIL) backend](https://github.com/triton-inference-server/fil_backend) on [Amazon Elastic Kubernetes Service (EKS)](https://aws.amazon.com/eks). [FIL](https://docs.rapids.ai/api/cuml/stable/api.html?highlight=forestinference#cuml.ForestInference) allows GPU accelerated inference for random forest and boosted decision tree models, and the FIL backend is right now available as a fully integrated part of Triton. We will use [Prometheus](https://prometheus.io/), [Prometheus Adapter](https://github.com/kubernetes-sigs/prometheus-adapter) and [dcgm-exporter](https://github.com/NVIDIA/gpu-monitoring-tools) (also known as NVIDIA GPU Monitoring Tools) to set up a Horizontal Pod Autoscaler (HPA) in EKS so that we can scale up our infrastructure based on inferencing load on the Triton Server.
It is assumed that you have an existing account with sufficient compute and GPU quota, a correctly configured `kubectl` utility, and a properly configured and accessible AWS S3 bucket.
**NOTE:** When referring to configuration parameters or elements that are `specific to your environment`, they will be represented as linux-style environment variables ex. ${YOUR_CONFIG_PARAM}.
Specific parameters used here include:
- REGION_NAME= < your preferred location >
- EKS_CLUSTER_NAME= < your cluster name >
- INSTANCE_TYPE= g4dn.12xlarge # or any other VM size with GPUs.
- AWS_ACCOUNT_ID= < select your ACCOUNT ID >
- REPOSITORY_NAME= titon_fil_backend # < name of the custom image we are going to build for triton backend, you can change this to your preference >
- S3_BUCKET_PATH= < s3 bucket name (existing or to be created) to host the models >
---
### Step 1: Prerequisites
- AWS account with at least the following permissions mentioned in https://eksctl.io/usage/minimum-iam-policies/ along with permissions to create, pull and push to AWS ECR repositories.
- System Softwares
- `docker`
- `aws-cli` , `eksctl`
- `helm` , `kubectl` , `istioctl`
- Python Libraries
```shell
pip install nvidia-pyindex
pip install tritonclient[all]
```
- Login credentials for aws-cli and kubectl configured.
**Note:** The rest of the demo uses `aws-cli` [version 2](https://docs.aws.amazon.com/cli/latest/userguide/install-cliv2.html). Hence some commands will be a little different from `aws-cli` version 1. We will also be using `helm` 3 that does not need `tiller`.
### Step 2: Obtain the Triton Inference Server official NVIDIA image (and optionally build a custom one)
The [FIL backend plugin](https://github.com/triton-inference-server/fil_backend) is natively supported in the official [Triton Inference Server](https://ngc.nvidia.com/catalog/containers/nvidia:tritonserver) image available in the [NVIDIA NGC Catalog](https://ngc.nvidia.com/catalog). The images in this catalogue are the preferred way of deploying the FIL backend. For this example, we will use the most current image: `nvcr.io/nvidia/tritonserver:21.06.1-py3`.
**You should skip to [**Step 3**](#Step-3:-Create-some-Triton-model-repository-entries,-or-use-the-provided-examples) if you are using the official NVIDIA image.**
For informational purpose, we will also show how to create a custom image with FIL backend below.
#### Step 2.a: Create an ECR repository
```shell
aws ecr get-login-password --region ${REGION_NAME} | docker login --username AWS --password-stdin ${AWS_ACCOUNT_ID}.dkr.ecr.${REGION_NAME}.amazonaws.com
aws ecr create-repository \
--repository-name ${REPOSITORY_NAME} \
--image-scanning-configuration scanOnPush=true \
--region ${REGION_NAME}
```
#### Step 2.b: Create and push the image to the ECR repository
**On a different terminal** do the following:
```shell
git clone https://github.com/triton-inference-server/fil_backend.git
cd fil_backend
docker build --tag ${AWS_ACCOUNT_ID}.dkr.ecr.${REGION_NAME}.amazonaws.com/${REPOSITORY_NAME} -f ops/Dockerfile .
docker push ${AWS_ACCOUNT_ID}.dkr.ecr.${REGION_NAME}.amazonaws.com/${REPOSITORY_NAME}:latest
```
### Step 3: Create some Triton model repository entries, or use the provided examples
A set of sample models, along with their `.pbtext` definition are provided in `./model_repository`. The layout structure and requirements are defined in the [Triton server docs](https://github.com/triton-inference-server/server/blob/master/docs/model_configuration.md),
and a brief introduction is also provided with the [FIL backend implementation](https://github.com/triton-inference-server/fil_backend/blob/main/README.md).
For the purpose of this demo we will use the models in `.model_repository` directory; however, feel free to add additional models. Make sure they follow the same structure in the docs. Here we assume the AWS `S3_BUCKET_PATH` bucket exists. You can create one in case it does not exist.
```shell
aws s3 cp --recursive ./model_repository s3://${S3_BUCKET_PATH}/model_repository/
aws s3 ls s3://${S3_BUCKET_PATH}/model_repository/
```
### Step 4: Configure an EKS cluster
We first need to create a Kubernetes cluster that will host our
Triton service, and has a GPU node pool with enough nodes and GPUs to illustrate horizontal scaling.
Lets first create a minimal yaml configuration file with name `eksctl_config.yaml` with the following in it
```yaml
apiVersion: eksctl.io/v1alpha5
kind: ClusterConfig
metadata:
name: <name set in EKS_CLUSTER_NAME>
region: <region set in REGION_NAME>
version: "1.20"
managedNodeGroups:
- name: nodegroup-managed-gpu-dev
minSize: 0
desiredCapacity: 2
maxSize: 4
instanceType: <instance type set in INSTANCE_TYPE>
ssh:
allow: true
```
To make things simple, we will use a [managed node group](https://docs.aws.amazon.com/eks/latest/userguide/managed-node-groups.html) to let EKS perform the provisioning and lifecycle management on our behalf. We are also allowing `ssh` access into the nodes for debugging. By default this will use the `~/.ssh/id_rsa.pub` file in your local machine.
With the `yaml` file, creating the EKS cluster is as simple as running the following command:
```shell
eksctl create cluster -f eksctl_config.yaml
```
This will take a few minutes before it completes. Grab a coffee :coffee: :coffee: . For more `yaml` configurations, you can refer to [eksctl documentation](https://eksctl.io/usage/schema/).
You can check whether the cluster is successfully created with following:
```shell
eksctl get cluster --name $EKS_CLUSTER_NAME --region $REGION_NAME
```
**NOTE:** For more `eksctl` related references on how to delete or modify the cluster, visit https://docs.aws.amazon.com/eks/latest/userguide/getting-started-eksctl.html.
Once the cluster is created successfully, let's get the credentials for your EKS cluster to access it from your machine.
```shell
aws eks update-kubeconfig --name $EKS_CLUSTER_NAME --region $REGION_NAME
```
Check whether you are able to access the nodes:
```shell
>> kubectl get nodes
NAME STATUS ROLES AGE VERSION
ip-172-31-12-100.ec2.internal NotReady <none> 10m v1.20.4-eks-6b7464
ip-172-31-34-168.ec2.internal NotReady <none> 10m v1.20.4-eks-6b7464
```
### Step 5: Setting up the EKS cluster to use GPUs for our workload
The good thing about using `eksctl` is that we can simply use a GPU compatible VM instance type with EKS and the AWS AMI resolvers will automatically select the correct EKS optimized accelerated AMI instance. Subsequently, `eksctl` will install the NVIDIA Kubernetes device plugin automatically ([reference](https://eksctl.io/usage/gpu-support/)) in the VMs. Therefore we do not have to do anything additional in this step.
### Step 6: To access models from AWS S3 and to fetch fil_backend image from ECR repository
To fetch the images from the ECR repository and to load the models from the AWS S3 we will use a config file and a secret. You need to convert the AWS credentials of your account in the base64 format and add it to the `./helm/charts/triton/values.yaml`, [similar to these instructions](https://github.com/triton-inference-server/server/tree/main/deploy/aws).
```shell
echo -n 'AWS_REGION' | base64
echo -n 'AWS_SECRET_KEY_ID' | base64
echo -n 'AWS_SECRET_ACCESS_KEY' | base64
echo -n 'AWS_SESSION_TOKEN' | base64
```
where `AWS_REGION`, `AWS_SECRET_KEY_ID`, `AWS_SECRET_ACCESS_KEY` and `AWS_SESSION_TOKEN` can be obtained either from the environment variable or the aws credentials file. `helm` will populate the `helm/charts/triton/templates/secrets.yaml` appropriately during deployment.
Replace the following fields in the `helm/charts/triton/values.yml` file with the above details
```yaml
image:
imageName: < update this with the image name you just pushed >
pullPolicy: IfNotPresent
modelRepositoryPath: s3://S3_BUCKET_PATH/model_repository
numGpus: 1
logVerboseLevel: 0
allowGPUMetrics: True
secret:
region: < replace with base64 AWS_REGION >
id: < replace with base64 AWS_SECRET_KEY_ID >
key: < replace with base64 AWS_SECRET_ACCESS_KEY >
session_token: < replace with base64 AWS_SESSION_TOKEN >
```
### Step 7: Install `istio` on your EKS cluster
Install the demo profile of `istio` by doing
```shell
istioctl install --set profile=demo
```
This will install the "Istio core", "Istiod", "Ingress gateways" and "Egress gateways" along with exposing all the necessary ports for gRPC and http connections for the ingress gateways.
Finally, enable sidecar injection to the triton pods in the default namespace:
```shell
kubectl label namespace default istio-injection=enabled
```
### Step 8: Configure Autoscaling
Autoscaling can be achieved in several tiers, pod autoscaling, node autoscaling etc. In this example we will set up a [Horizontal Pod Autoscaler (HPA)](https://docs.aws.amazon.com/eks/latest/userguide/horizontal-pod-autoscaler.html) that autoscales pods based on custom metrics obtained from a custom service [dcgm-exporter](https://github.com/NVIDIA/gpu-monitoring-tools) , also known as NVIDIA GPU Monitoring Tools.
With increasing inferencing load in the FIL inferencing system, the GPU utilization will go up. As a result, the SLAs may be hampered. If we have a HPA in place, our cluster can then horizontally scale up to add more pods with GPU access to handle this additional load.
##### We will need to setup a few things in order for this setup to work:
- [nvidia-device-plugin](https://github.com/NVIDIA/k8s-device-plugin): We already have the NVIDIA k8s device plugin installed automatically by AWS.
- [kube-prometheus-stack](https://github.com/prometheus-community/helm-charts/tree/main/charts/kube-prometheus-stack) - This collects the GPU metrics, store them and display in Grafana dashboard.
- [prometheus-adapter](https://github.com/kubernetes-sigs/prometheus-adapter) - Collects metrics from the Prometheus server and exposes them to a custom k8s metrics api for use in pod autoscaling.
- [dcgm-exporter](https://github.com/NVIDIA/gpu-monitoring-tools) - Esentially a daemonset with some services to reveal GPU metrics on each node.
#### Step 8.a: Install Prometheus related components
Installing the Prometheus stack is made easy with the help of `Helm` charts. We will first add the Prometheus kube stack Helm chart to our local Helm installation. We then make some customizations before installing it. In addition, we will also install the prometheus-adapter stack.
Do the following to get the default configuration values of the kube-prometheus-stack chart in a `yaml` file to modify the settings.
```shell
helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
helm repo update
helm inspect values prometheus-community/kube-prometheus-stack > ./kube-prometheus-stack.values.yaml
```
Next, we’ll need to edit the values file to change the port at which the Prometheus server service is available. In the `prometheus` instance section of the chart yaml file, change the service type from ClusterIP to NodePort. This will allow the Prometheus server to be accessible at your machine ip address at port 30090 as http://<machine-ip>:30090/ using `kubectl port-forward` if you are on a remote machine.
From:
```yaml
# Port to expose on each node
# Only used if service.type is 'NodePort'
#
nodePort: 30090
# Loadbalancer IP
# Only use if service.type is "loadbalancer"
loadBalancerIP: ""
loadBalancerSourceRanges: []
# Service type
#
type: ClusterIP
```
To:
```yaml
# Port to expose on each node
# Only used if service.type is 'NodePort'
#
nodePort: 30090
# Loadbalancer IP
# Only use if service.type is "loadbalancer"
loadBalancerIP: ""
loadBalancerSourceRanges: []
# Service type
#
type: NodePort
```
Also, modify the `prometheusSpec.serviceMonitorSelectorNilUsesHelmValues` settings to `false` below to allow Prometheus to scrape metrics from all namespaces :
```yaml
# If true, a nil or {} value for prometheus.prometheusSpec.serviceMonitorSelector will cause the
# prometheus resource to be created with selectors based on values in the helm deployment,
# which will also match the servicemonitors created
#
serviceMonitorSelectorNilUsesHelmValues: false
```
Lastly, we also add the following configMap to the section on `additionalScrapeConfigs` in the Helm chart. Make sure that in the `additionalScrapeConfigs.kubernetes_sd_configs.role.namespaces.names` you specify the namespace where the `dcgm-exporter` service will be deployed so that Prometheus can scrape metrics information from that service. Here we will deploy the dcgm exporter in the `default` namespace, so we keep the value `default`:
```yaml
# AdditionalScrapeConfigs allows specifying additional Prometheus scrape configurations. Scrape configurations
# are appended to the configurations generated by the Prometheus Operator. Job configurations must have the form
# as specified in the official Prometheus documentation:
# https://prometheus.io/docs/prometheus/latest/configuration/configuration/#scrape_config. As scrape configs are
# appended, the user is responsible to make sure it is valid. Note that using this feature may expose the possibility
# to break upgrades of Prometheus. It is advised to review Prometheus release notes to ensure that no incompatible
# scrape configs are going to break Prometheus after the upgrade.
#
# The scrape configuration example below will find master nodes, provided they have the name .*mst.*, relabel the
# port to 2379 and allow etcd scraping provided it is running on all Kubernetes master nodes
#
additionalScrapeConfigs:
- job_name: gpu-metrics
scrape_interval: 1s
metrics_path: /metrics
scheme: http
kubernetes_sd_configs:
- role: endpoints
namespaces:
names:
- default
# - <the namespace where dcgm exporter will be installed. Needs to be in the same namespace.>
relabel_configs:
- source_labels: [__meta_kubernetes_pod_node_name]
action: replace
target_label: kubernetes_node
```
Next, we will finally install the Prometheus stack and the Prometheus adapter.
```shell
# Install Prometheus stack with the modifications
helm install prometheus-community/kube-prometheus-stack \
--create-namespace --namespace prometheus \
--generate-name \
--values ./kube-prometheus-stack.values.yaml
# Get the Prometheus Service
PROM_SERVICE=$(kubectl get svc -nprometheus -lapp=kube-prometheus-stack-prometheus -ojsonpath='{range .items[*]}{.metadata.name}{"\n"}{end}')
# Install Prometheus Adapter
helm install prometheus-adapter prometheus-community/prometheus-adapter \
--namespace prometheus \
--set rbac.create=true,prometheus.url=http://${PROM_SERVICE}.prometheus.svc.cluster.local,prometheus.port=9090
```
Once you install the above charts, in around 1-2minutes, you should be able to observe metrics from the custom endpoint
```shell
kubectl get --raw /apis/custom.metrics.k8s.io/v1beta1 | jq -r .
```
### Step 9: Deploy the dcgm exporter stack.
**Note 1:** It is mandatory that the Prometheus stack is deployed beforehand.
**Note 2:** The instructions to configure dcgm are modified from [official NVIDIA instructions](https://docs.nvidia.com/datacenter/cloud-native/kubernetes/dcgme2e.html#gpu-telemetry) .
The dcgm-exporter as of now may run into a few problems when deploying. Fortunately, we have some gotchas. We will have to make sure we customize our dcgm deployment to take care of those.
- Firstly, in the current version of dcgm-exporter, the GPU utilization metric `DCGM_FI_DEV_GPU_UTIL` is turned off by default, probably because it is resource intensive to collect (see https://github.com/NVIDIA/gpu-monitoring-tools/issues/143). We need to enable this flag to obtain GPU utilization metrics for autoscaling.
- Secondly, there is a liveness bug where the pods after deployment will consistently enter a CrashLoopBackoff status (https://github.com/NVIDIA/gpu-monitoring-tools/issues/120). To get around this, we need to modify the dcgm-exporter k8s config by taking details from [here](https://github.com/NVIDIA/gpu-monitoring-tools/issues/120#issuecomment-801574290) and customize them slightly.
To fix the first problem and re-enable the metric `DCGM_FI_DEV_GPU_UTIL`, we will create a custom docker image and use that image in the dcgm-exporter deployment. Then to mitigate the second issue, we will modify the helm chart of dcgm-exporter before deploying.
First let us create a docker image with the following Dockerfile to expose our required metrics:
```Dockerfile
FROM nvcr.io/nvidia/k8s/dcgm-exporter:latest
RUN sed -i -e '/^# DCGM_FI_DEV_GPU_UTIL.*/s/^#\ //' /etc/dcgm-exporter/default-counters.csv
ENTRYPOINT ["/usr/local/dcgm/dcgm-exporter-entrypoint.sh"]
```
**Note:** You can choose to expose any other metric of interest in the above Dockerfile as well.
Let's suppose, you build and push the image to [Dockerhub](https://hub.docker.com/) as `anirbandas/dcgm-exporter:latest`.
Next, clone the [NVIDIA gpu-monitoring-tools](https://github.com/NVIDIA/gpu-monitoring-tools) repository which contains the dcgm-exporter.
```shell
git clone https://github.com/NVIDIA/gpu-monitoring-tools
```
Then edit the `./gpu-monitoring-tools/deployment/dcgm-exporter/values.yaml` file
From :
```yaml
image:
repository: nvcr.io/nvidia/k8s/dcgm-exporter
pullPolicy: IfNotPresent
# Image tag defaults to AppVersion, but you can use the tag key
# for the image tag, e.g:
tag: 2.1.8-2.4.0-rc.3-ubuntu18.04
# Comment the following line to stop profiling metrics from DCGM
arguments: ["-f", "/etc/dcgm-exporter/dcp-metrics-included.csv"]
```
To:
```yaml
image:
repository: anirbandas/dcgm-explorer # or whereever you pushed your image
pullPolicy: IfNotPresent
# Image tag defaults to AppVersion, but you can use the tag key
# for the image tag, e.g:
tag: latest
# The following line will use the default metrics flags
arguments: ["-k"]
```
Also, do not forget to edit the `livenessProbe` and `readinessProbe` fields of the daemonset configuration at `./gpu-monitoring-tools/deployment/dcgm-exporter/templates/daemonset.yaml` to get rid of the `CrashLoopBackoff` error. We change the config as follows.
From :
```yaml
livenessProbe:
httpGet:
path: /health
port: {{ .Values.service.port }}
initialDelaySeconds: 5
periodSeconds: 5
readinessProbe:
httpGet:
path: /health
port: {{ .Values.service.port }}
initialDelaySeconds: 5
```
To:
```yaml
livenessProbe:
httpGet:
path: /health
port: {{ .Values.service.port }}
initialDelaySeconds: 20
periodSeconds: 5
readinessProbe:
httpGet:
path: /health
port: {{ .Values.service.port }}
initialDelaySeconds: 20
```
You should then be able to deploy the dcgm-exporter as:
```shell
helm install dcgm-exporter gpu-monitoring-tools/deployment/dcgm-exporter/
```
To check whether the GPU metrics are available in the Prometheus custom metrics api endpoint:
```shell
kubectl get --raw /apis/custom.metrics.k8s.io/v1beta1 | jq -r . | grep DCGM_FI_DEV_GPU_UTIL
```
You want to be able to see something like:
```shell
"name": "namespaces/DCGM_FI_DEV_GPU_UTIL",
"name": "services/DCGM_FI_DEV_GPU_UTIL",
"name": "jobs.batch/DCGM_FI_DEV_GPU_UTIL",
"name": "pods/DCGM_FI_DEV_GPU_UTIL",
```
If you are upto this point, then you are golden. You can now pull the custom metrics from both Prometheus and Grafana dashboards.
### Step 10: Configure the HPA
Now that everything else is in place, and we have the GPU metrics available in the custom metrics server, we will configure our HPA before we deploy the Triton service.
Our `hpa` configuration file is located at [./helm/charts/triton/templates/hpa.yml](./helm/charts/triton/templates/hpa.yml), and it should look like the following before deployment.
```yaml
apiVersion: autoscaling/v2beta1
kind: HorizontalPodAutoscaler
metadata:
name: triton-hpa
namespace: {{ .Release.Namespace }}
labels:
app: triton-hpa
spec:
minReplicas: {{ .Values.minReplicaCount }}
maxReplicas: {{ .Values.maxReplicaCount }}
metrics:
- type: Object
object:
metricName: DCGM_FI_DEV_GPU_UTIL
targetValue: {{ .Values.HPATargetAverageValue }}
target:
kind: Service
name: {{ .Values.DCGMExporterService }}
scaleTargetRef:
apiVersion: apps/v1
kind: Deployment
name: {{ template "triton-inference-server.name" . }}
```
Update `./helm/charts/triton/Values.yaml` with the correct dcgm-exporter service name as
```yaml
.
.
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
initReplicaCount: 1
minReplicaCount: 2
maxReplicaCount: 6
HPATargetAverageValue: 80
DCGMExporterService: dcgm-exporter # <replace with the DCGM exporter service name if different>
image:
.
.
```
### Step 11: Deploy the Triton Inferencing Server
Once everything else is configured, deploying triton service is as easy as the following:
```shell
helm install triton ./helm/charts/triton/
```
### Step 12: Check if the HPA is able to access the metrics
It is a good idea to check if the HPA is able to get the metrics from the metrics server. If there is an error the HPA would not autoscale your pods.
```shell
kubectl describe hpa triton-hpa
```
If everything else is configured correctly, you will be able to see something similar to the following:
```shell
Name: triton-hpa
Namespace: default
Labels: app=triton-hpa
app.kubernetes.io/managed-by=Helm
Annotations: meta.helm.sh/release-name: triton
meta.helm.sh/release-namespace: default
CreationTimestamp: Wed, 07 Jul 2021 18:43:06 -0400
Reference: Deployment/triton-inference-server
Metrics: ( current / target )
"DCGM_FI_DEV_GPU_UTIL" on Service/dcgm-exporter-1625697000 (target value): 0 / 800m
Min replicas: 2
Max replicas: 6
Deployment pods: 2 current / 2 desired
Conditions:
Type Status Reason Message
---- ------ ------ -------
AbleToScale True ScaleDownStabilized recent recommendations were higher than current one, applying the highest recent recommendation
ScalingActive True ValidMetricFound the HPA was able to successfully calculate a replica count from Service metric DCGM_FI_DEV_GPU_UTIL
ScalingLimited False DesiredWithinRange the desired count is within the acceptable range
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal SuccessfulRescale 3m3s horizontal-pod-autoscaler New size: 2; reason: Current number of replicas below Spec.MinReplicas
```
### Step 12: Exploring inference
At this point, your triton inference cluster should be up and running or in process of coming up. Now we can submit some test data to our running server. The process for doing this, assuming the default model, is illustrated in the jupyter notebook [triton_inference.ipynb](./triton_inference.ipynb).
### Step 13. For cluster autoscaling you can follow instructions in https://github.com/awsdocs/amazon-eks-user-guide/blob/master/doc_source/cluster-autoscaler.md .
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/model_repository | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/model_repository/cuml_model/config.pbtxt | name: "cuml_model"
backend: "fil"
max_batch_size: 8192
input [
{
name: "input__0"
data_type: TYPE_FP32
dims: [ 32 ]
}
]
output [
{
name: "output__0"
data_type: TYPE_FP32
dims: [ 1 ]
}
]
instance_group [{ kind: KIND_GPU }]
parameters [
{
key: "model_type"
value: { string_value: "treelite_checkpoint" }
},
{
key: "predict_proba"
value: { string_value: "false" }
},
{
key: "output_class"
value: { string_value: "true" }
},
{
key: "threshold"
value: { string_value: "0.5" }
},
{
key: "algo"
value: { string_value: "ALGO_AUTO" }
},
{
key: "storage_type"
value: { string_value: "AUTO" }
},
{
key: "blocks_per_sm"
value: { string_value: "0" }
}
]
dynamic_batching {
preferred_batch_size: [1, 2, 4, 8, 16, 32, 64, 128, 1024, 2048, 4096, 8192]
max_queue_delay_microseconds: 30000
} | 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/model_repository | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/model_repository/xgb_model/config.pbtxt | name: "xgb_model"
backend: "fil"
max_batch_size: 8192
input [
{
name: "input__0"
data_type: TYPE_FP32
dims: [ 32 ]
}
]
output [
{
name: "output__0"
data_type: TYPE_FP32
dims: [ 1 ]
}
]
instance_group [{ kind: KIND_GPU }]
parameters [
{
key: "model_type"
value: { string_value: "xgboost" }
},
{
key: "predict_proba"
value: { string_value: "false" }
},
{
key: "output_class"
value: { string_value: "true" }
},
{
key: "threshold"
value: { string_value: "0.5" }
},
{
key: "algo"
value: { string_value: "ALGO_AUTO" }
},
{
key: "storage_type"
value: { string_value: "AUTO" }
},
{
key: "blocks_per_sm"
value: { string_value: "0" }
}
]
dynamic_batching {
preferred_batch_size: [1, 2, 4, 8, 16, 32, 64, 128, 1024, 2048, 4096, 8192]
max_queue_delay_microseconds: 30000
} | 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/model_repository | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/model_repository/light_model/config.pbtxt | name: "light_model"
backend: "fil"
max_batch_size: 8192
input [
{
name: "input__0"
data_type: TYPE_FP32
dims: [ 32 ]
}
]
output [
{
name: "output__0"
data_type: TYPE_FP32
dims: [ 1 ]
}
]
instance_group [{ kind: KIND_GPU }]
parameters [
{
key: "model_type"
value: { string_value: "lightgbm" }
},
{
key: "predict_proba"
value: { string_value: "false" }
},
{
key: "output_class"
value: { string_value: "true" }
},
{
key: "threshold"
value: { string_value: "0.5" }
},
{
key: "algo"
value: { string_value: "ALGO_AUTO" }
},
{
key: "storage_type"
value: { string_value: "AUTO" }
},
{
key: "blocks_per_sm"
value: { string_value: "0" }
}
]
dynamic_batching {
preferred_batch_size: [1, 2, 4, 8, 16, 32, 64, 128, 1024, 2048, 4096, 8192]
max_queue_delay_microseconds: 30000
} | 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/model_repository/light_model | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/model_repository/light_model/1/model.txt | tree
version=v3
num_class=1
num_tree_per_iteration=1
label_index=0
max_feature_idx=31
objective=binary sigmoid:1
feature_names=Column_0 Column_1 Column_2 Column_3 Column_4 Column_5 Column_6 Column_7 Column_8 Column_9 Column_10 Column_11 Column_12 Column_13 Column_14 Column_15 Column_16 Column_17 Column_18 Column_19 Column_20 Column_21 Column_22 Column_23 Column_24 Column_25 Column_26 Column_27 Column_28 Column_29 Column_30 Column_31
feature_infos=[-6.5546088218688965:6.2278804779052734] [-7.0495986938476562:5.7515859603881836] [-4.965395450592041:6.5972399711608887] [-4.6478652954101562:6.5585684776306152] [-8.2879343032836914:7.7419576644897461] [-6.8546819686889648:7.0196652412414551] [-8.7973241806030273:5.5609464645385742] [-6.5725240707397461:7.2596492767333984] [-6.2245054244995117:7.497591495513916] [-6.8401269912719727:6.9245181083679199] [-11.817563056945801:12.412569046020508] [-14.371319770812988:12.11441707611084] [-2.6720693111419678:3.1814792156219482] [-3.1954507827758789:3.3344674110412598] [-3.2284896373748779:3.2031958103179932] [-4.0048222541809082:3.3795340061187744] [-3.2715401649475098:2.9391093254089355] [-3.5124735832214355:2.8368351459503174] [-2.8929731845855713:3.4158065319061279] [-3.7133853435516357:3.583622932434082] [-2.697335958480835:3.3189976215362549] [-2.9512476921081543:3.3829400539398193] [-3.5988059043884277:2.945188045501709] [-3.780109167098999:2.9388504028320312] [-3.300701379776001:3.0906422138214111] [-3.132082462310791:3.1528604030609131] [-2.9600803852081299:3.1746981143951416] [-3.3740308284759521:3.2765915393829346] [-2.8621206283569336:3.2689967155456543] [-3.3815498352050781:2.7025766372680664] [-3.0758681297302246:3.1015768051147461] [-3.216130256652832:2.6690890789031982]
tree_sizes=3342 3332 3343 3340 3334 3339 3350 3345 3332 3342 3359 3345 3360 3366 3343 3365 3358 3358 3342 3359 3347 3342 3356 3353 3348 3362 3348 3362 3371 3349 3372 3356 3371 3355 3381 3368 3387 3371 3378 3385 3381 3367 3371 3375 3362 3386 3390 3398 3398 3394 3405 3401 3402 3411 3390 3420 3414 3422 3428 3430 3419 3416 3417 3424 3432 3427 3427 3434 3425 3420 3411 3441 3425 3442 3431 3436 3445 3441 3439 3437 3437 3446 3466 3448 3475 3453 3452 3453 3460 3460 3472 3458 3475 3456 3475 3466 3494 3474 3480 3480
Tree=0
num_leaves=31
num_cat=0
split_feature=8 3 0 4 4 7 9 4 0 8 0 1 25 5 9 1 10 24 6 8 3 7 20 11 6 14 22 20 1 24
split_gain=171.581 98.365 68.6383 29.8722 27.3854 20.2096 15.8694 15.5581 13.7042 28.3754 11.6834 13.4982 13.3367 10.0874 14.2514 9.32147 11.0366 6.40164 6.24569 5.92079 6.32015 7.40421 5.81967 3.09674 2.92103 2.6751 2.3178 1.67184 0.963037 0.578417
threshold=0.81758242845535289 -0.70274627208709706 -1.7083070278167722 -1.6327362060546873 0.45788866281509405 -0.17649734020233152 -0.13153135776519773 -0.12080970406532286 0.63517421483993541 -0.87626728415489186 -1.1507846117019651 0.30255699157714849 -0.64878144860267628 -0.89550951123237599 0.023247927427291874 -1.4832193851470945 0.98919457197189342 -0.29334968328475947 -0.65812426805496205 -0.62308597564697255 1.4703543782234194 -0.20904585719108579 -0.33045321702957148 -2.946575284004211 -0.37206447124481196 -0.2194641828536987 0.19348140060901645 0.59053277969360363 -0.27431529760360712 1.2365222573280337
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 3 7 6 19 15 -1 -3 26 -10 -2 12 17 24 23 -6 -17 -12 -13 29 21 -21 -11 -15 -14 -9 -5 -20 -7 -4
right_child=10 2 4 8 5 28 -8 25 9 22 11 18 13 14 -16 16 -18 -19 27 20 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.059579021832063955 -0.21758263937078751 0.031213454693273864 0.17124930831710006 -0.2140345217590354 0.075224722975047731 0.11123394247831739 0.17124930831710011 -0.12882752087681318 0.13123906442457831 -0.22885313060811757 -0.068812155038030509 -0.10990375687359338 -0.22885313060811754 -0.19074813642476346 0.062130461337495245 -0.1514139488806561 0.027212430304021706 0.091228820532056554 -0.22885313060811752 -0.020799862367004405 0.16149071224575329 0.13645779188882021 -0.08336133463531116 -0.083361334635311146 -0.13652179854845192 -0.22885313060811763 -0.14311689369557093 -0.16729890923500715 0.17124930831710011 0.13790743840666528
leaf_weight=6.498335987329483 17.745455965399742 4.9987199902534494 39.239951923489571 13.496543973684313 6.248399987816815 4.9987199902534485 5.4985919892787916 4.9987199902534476 4.9987199902534476 5.4985919892787924 4.9987199902534485 9.2476319819688815 7.2481439858675003 5.2486559897661218 5.4985919892787924 7.7480159848928452 6.2483999878168106 4.9987199902534467 13.746479973196985 6.2483999878168106 10.247375980019568 5.7485279887914675 5.4985919892787924 5.4985919892787942 6.4983359873294848 5.7485279887914631 6.998207986354827 6.4983359873294821 5.748527988791464 5.9984639883041373
leaf_count=26 71 20 157 54 25 20 22 20 20 22 20 37 29 21 22 31 25 20 55 25 41 23 22 22 26 23 28 26 23 24
internal_value=-0.0320027 0.0286629 0.0793221 -0.0906734 0.11032 0.0325041 0.0462173 -0.114538 -0.135679 -0.0663115 -0.145164 -0.126669 -0.0888173 -0.122159 -0.0688122 -0.0263321 -0.07167 0.0112083 -0.177993 0.146058 0.103816 0.0545528 -0.156107 -0.135806 -0.185206 -0.18233 -0.189819 -0.209095 0.143335 0.166828
internal_weight=0 162.708 114.221 48.4876 98.4748 30.9921 11.9969 15.746 36.4907 15.9959 87.2277 69.4822 39.9898 29.9923 16.2458 20.2448 13.9964 9.99744 29.4924 67.4827 22.2443 11.9969 10.9972 10.7472 13.7465 10.7472 20.4948 20.2448 10.7472 45.2384
internal_count=1000 651 457 194 394 124 48 63 146 64 349 278 160 120 65 81 56 40 118 270 89 48 44 43 55 43 82 81 43 181
is_linear=0
shrinkage=1
Tree=1
num_leaves=31
num_cat=0
split_feature=8 3 0 4 4 7 8 0 8 0 1 0 9 2 6 10 7 15 20 8 3 9 12 5 2 26 6 25 18 12
split_gain=140.008 80.1301 56.5356 24.6214 24.3925 16.1956 14.9208 11.2424 23.128 10.0612 15.5656 9.03321 8.95813 7.61489 6.96973 13.5742 12.8454 5.30388 4.8467 3.84585 6.13997 9.02092 3.69914 3.004 2.88322 2.18884 3.15578 2.13685 3.0624 1.16727
threshold=0.81758242845535289 -0.70274627208709706 -1.8935308456420896 0.14497005939483645 -1.6327362060546873 -0.20904585719108579 -0.82658091187477101 0.63517421483993541 -0.87626728415489186 -0.53437867760658253 -2.1899367570877071 1.5746666789054873 2.2621223926544194 1.4999915361404421 0.9321291744709016 0.22259239107370379 0.72382399439811718 -0.57546964287757862 -0.33045321702957148 -0.62308597564697255 1.4703543782234194 -0.76543956995010365 -0.27862833440303797 -0.25183287262916559 2.3047043085098271 0.3114320188760758 -1.2840110063552854 0.36171150207519537 0.031244269572198394 0.62888491153717052
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 4 13 19 6 11 -1 27 -9 25 -11 12 22 -3 15 16 17 -12 -10 -4 21 -21 -5 -7 -16 -2 -27 28 -6 -19
right_child=9 2 3 5 7 23 -8 8 18 10 14 -13 -14 -15 24 -17 -18 29 -20 20 -22 -23 -24 -25 -26 26 -28 -29 -30 -31
leaf_value=0.19224933079982384 -0.18178147676044246 -0.016645444459515258 0.18037690834842077 -0.1362490514518207 -0.16600349181842156 0.08789764108061493 -0.031993587114762988 0.14752832429651616 -0.17954453554248911 0.037301643419057624 -0.051563907697959988 0.11668319335775784 0.083719941792758115 -0.17030824887348617 -0.1845662205469053 0.067959992225334026 0.0096447725562566911 -0.18662120592872855 -0.046294481557376706 -0.022664721932374805 0.18508744590406764 0.15940003114512158 -0.025002212949256038 0.18640986050495617 -0.097133482478666378 -0.18288780805587426 -0.076703180668533266 -0.18289826158773406 -0.065516069460689177 -0.12970203109642495
leaf_weight=5.4726904928684235 14.091684758663179 6.2260060310363752 43.206142380833626 5.2279839813709277 6.1883575469255465 4.9709692150354376 6.4814170598983747 4.9785318970680228 5.4286100864410392 9.4589523077011091 4.9745479524135616 5.9829215854406348 5.2350328862667075 6.6905955523252487 11.157539337873461 5.4696031808853141 7.4618033468723288 11.672343090176588 5.4904560148715964 4.9926604330539703 9.9350839853286725 5.9823538810014742 6.9792296439409292 8.2041519433259946 5.6976270228624335 4.948859766125679 6.4424685984849912 8.1750198751687986 5.9474616646766663 5.2115984112024298
leaf_count=22 57 25 174 21 25 20 26 20 22 38 20 24 21 27 45 22 30 47 22 20 40 24 28 33 23 20 26 33 24 21
internal_value=0 0.0549067 0.100723 0.125983 -0.05319 0.0605419 0.0706667 -0.0940809 -0.0310978 -0.102646 -0.0806328 0.0106543 -0.025715 -0.0962403 -0.102233 -0.0766627 -0.103642 -0.142314 -0.112542 0.163339 0.128134 0.0765767 -0.0726459 0.149241 -0.155011 -0.155431 -0.122834 -0.143379 -0.116757 -0.169052
internal_weight=0 161.796 113.633 100.717 48.1625 36.6003 11.9541 36.2084 15.8976 86.587 61.104 23.4252 17.4422 12.9166 51.6451 34.7899 29.3203 21.8585 10.9191 64.1162 20.9101 10.975 12.2072 13.1751 16.8552 25.483 11.3913 20.3108 12.1358 16.8839
internal_count=1000 651 457 405 194 147 48 146 64 349 246 94 70 52 208 140 118 88 44 258 84 44 49 53 68 103 46 82 49 68
is_linear=0
shrinkage=0.1
Tree=2
num_leaves=31
num_cat=0
split_feature=8 8 7 4 3 0 0 4 11 1 5 7 8 0 1 5 20 7 4 8 0 3 15 30 27 12 14 22 0 3
split_gain=117.061 44.4902 48.2537 37.8674 23.1903 16.2677 14.4783 10.9941 12.5451 12.8884 10.8488 9.24217 9.22051 8.05594 11.3227 12.2348 7.16251 5.81748 5.4689 4.20034 3.83261 6.03259 3.11742 1.8526 1.63233 1.56444 0.206249 0.194933 0.170674 0.0355673
threshold=-1.1515020728111265 1.1326235532760622 0.25074130296707159 0.74304640293121349 0.7560294270515443 0.23862367868423465 -0.87671247124671925 -2.4476122856140132 -0.38942228257656092 0.36228165030479437 0.29740685224533087 -2.1114803552627559 -0.043257951736450188 -0.53437867760658253 -0.31010645627975458 -0.29206293821334833 0.47881735861301428 0.72382399439811718 0.52956873178482067 2.5029540061950688 -0.43635436892509455 1.2886238098144533 -0.57546964287757862 -0.47790913283824915 0.04701561480760575 -0.6693873107433318 1.1974483132362368 -0.25551988184452051 0.8488212823867799 -0.9943212866783141
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=3 2 4 20 7 11 -6 -2 23 -10 12 -5 -4 25 15 19 24 22 -8 -15 21 -1 -16 -9 -11 -3 -22 -12 -24 -25
right_child=1 13 10 5 6 -7 18 8 9 16 27 -13 -14 14 17 -17 -18 -19 -20 -21 26 -23 28 29 -26 -27 -28 -29 -30 -31
leaf_value=0.01984787095863599 0.0417020575280087 -0.11290815517446716 0.15709080772019768 -0.16205312868015018 -0.077479354205590878 0.11201130512959732 0.14798985803882173 -0.11279605847572322 0.063841001064714395 -0.17530989311664646 0.15880964776997958 0.016528167363553163 -0.010666516628428909 -0.030912797975377354 -0.07676564740886116 0.058644165665683895 0.00011295205047962304 -0.037266603038700802 0.029734590312484679 -0.14230388271324748 0.17291562878069588 0.16228686327457909 -0.15118933267248985 -0.17154324406071944 -0.098384250976079654 -0.17048761373480958 0.15087350536099295 0.18370706746548587 -0.17330271770900132 -0.18191936754465704
leaf_weight=4.8982804715633375 6.4351709634065619 6.7873481512069729 5.1366530805826196 6.5911158919334421 7.1436536014080074 6.6647720932960501 8.5863281935453397 7.1010795831680324 7.1645105332136181 6.3415598422288912 4.8856654316186896 5.1721473485231408 9.0467385351657867 6.4256303608417529 4.9063057750463548 9.4362122863531095 5.2009715288877478 6.1603895872831336 7.1817438304424277 7.1542515456676483 33.556421399116516 7.566150426864624 4.8541645407676723 8.4835261106491089 4.8820448219776154 15.482323884963987 4.8598640561103812 8.8248396664857847 12.421933531761168 5.4103879183530781
leaf_count=20 26 28 21 27 29 27 35 29 29 26 20 21 37 26 20 38 21 25 29 29 138 31 20 35 20 64 20 36 51 22
internal_value=0 -0.0434878 -0.000667689 0.11001 -0.0429521 -0.0128117 0.0406233 -0.0804841 -0.0981202 -0.0480764 0.111404 -0.0835334 0.0500884 -0.102706 -0.0809246 -0.0288204 -0.0968955 -0.123236 0.0941292 -0.0895967 0.154494 0.106311 -0.147112 -0.154347 -0.141849 -0.152939 0.170127 0.174835 -0.167089 -0.175584
internal_weight=0 175.453 101.825 69.3088 73.931 18.428 22.9117 51.0193 44.5841 23.5891 27.8939 11.7633 14.1834 73.6286 51.3589 23.0161 16.4246 28.3428 15.7681 13.5799 50.8807 12.4644 22.1824 20.995 11.2236 22.2697 38.4163 13.7105 17.2761 13.8939
internal_count=1000 716 415 284 301 75 93 208 182 96 114 48 58 301 209 93 67 116 64 55 209 51 91 86 46 92 158 56 71 57
is_linear=0
shrinkage=0.1
Tree=3
num_leaves=31
num_cat=0
split_feature=8 8 3 4 0 0 7 10 1 10 3 7 20 0 9 5 7 9 18 19 0 3 6 29 3 5 24 13 11 9
split_gain=97.3853 36.9644 41.1266 32.0296 30.2502 13.3818 11.7056 10.5395 13.1813 9.73636 9.62122 7.8009 7.22479 7.10867 9.72983 15.0273 5.76034 5.3035 5.208 5.11298 3.44479 5.18154 3.20629 2.5539 6.60098 2.12341 1.75054 1.66114 0.692294 0.205851
threshold=-1.1515020728111265 1.1326235532760622 -0.9574308693408965 0.74304640293121349 -0.89208686351776112 0.23862367868423465 0.19911569356918338 0.29159404337406164 -1.8522490859031675 4.0049445629119882 0.90646380186080944 -2.1114803552627559 0.47881735861301428 -0.32855796813964838 0.57707139849662792 -0.85830512642860401 1.0554251670837405 -0.036320179700851433 0.022797232493758205 0.84618529677391063 -0.43635436892509455 1.2886238098144533 1.159727215766907 0.022604688070714477 -2.0779349803924556 -0.4936488270759582 0.57650366425514232 -0.53143158555030812 1.9473140835762026 2.193432211875916
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=3 2 12 20 9 11 7 8 -6 10 -4 -5 23 26 16 -16 19 -14 -10 27 21 -1 -9 24 -2 -8 29 -15 -29 -3
right_child=1 13 4 5 6 -7 25 22 18 -11 -12 -13 17 14 15 -17 -18 -19 -20 -21 -22 -23 -24 -25 -26 -27 -28 28 -30 -31
leaf_value=0.017865907761750632 -0.1557141807536655 -0.16163728098973584 -0.094212342087465886 -0.15019619730700395 0.12038250081889015 0.10204399477161424 0.097139183150793645 0.16039310631024184 -0.11731646051922731 -0.17552386442327578 0.085713080346222922 0.014867522784127602 0.049328744818053161 -0.090321880633872087 -0.088629839143961039 0.13889271573284845 -0.023724060521661192 -0.098570754952137185 0.010089845145560702 -0.042971381012413788 0.15870605812246244 0.1506557036785601 0.057747079120681104 -0.16715625778548818 -0.0062801545546637208 0.16643547730843791 -0.089338621599678369 -0.16917598359134128 -0.12508641306442178 -0.13716489438415624
leaf_weight=4.8975395411252958 7.633879482746126 13.160880580544477 7.0237592458724993 6.4094100594520578 5.9403825253248259 6.5880389213562003 6.7367670685052898 6.9989861845970136 7.0747662335634232 5.7982353270053855 5.1519139111042023 5.1747220009565362 4.9144538193941116 4.8344904780387905 7.0233998000621796 4.9479817003011686 8.1072786599397642 4.785413593053816 5.8708141446113586 6.7730698734521857 37.127679541707039 7.3463107943534851 5.384094625711441 9.2754390537738782 4.8240164220333135 12.869051247835158 5.208317205309867 12.453267276287084 4.9877932518720618 4.6521514952182761
leaf_count=20 32 56 29 27 24 27 28 29 29 24 21 21 20 20 29 20 33 20 24 28 158 31 22 39 20 54 22 52 21 20
internal_value=0 -0.0398951 -0.000622732 0.101759 0.0426482 -0.0117467 0.0820473 0.0440651 -0.00294572 -0.0688702 -0.0180802 -0.0764609 -0.0953999 -0.0944813 -0.0729938 0.00540922 -0.0982548 -0.0236372 -0.0595378 -0.119056 0.143537 0.0975398 0.115763 -0.127429 -0.0978495 0.142625 -0.140335 -0.14219 -0.156567 -0.155246
internal_weight=0 172.431 100.282 67.5437 68.8488 18.1722 50.8749 31.269 18.886 17.9739 12.1757 11.5841 31.4332 72.1486 49.1273 11.9714 37.1559 9.69987 12.9456 29.0486 49.3715 12.2439 12.3831 21.7333 12.4579 19.6058 23.0213 22.2756 17.4411 17.813
internal_count=1000 716 415 284 284 75 210 128 77 74 50 48 131 301 203 49 154 40 53 121 209 51 51 91 52 82 98 93 73 76
is_linear=0
shrinkage=0.1
Tree=4
num_leaves=31
num_cat=0
split_feature=8 4 8 7 3 0 9 0 11 5 9 6 6 7 6 7 5 4 4 20 0 1 9 12 27 3 3 5 3 0
split_gain=81.9559 35.5865 29.0444 35.6188 17.2191 16.4192 15.0376 14.0141 9.01091 6.63721 10.9731 10.4845 10.4808 8.61683 7.68887 6.4433 6.02714 5.5908 4.41165 3.90469 3.83201 2.54125 2.3012 2.0467 0.647 0.634531 0.608848 0.231048 0.0642142 0.0379159
threshold=-0.95728960633277882 0.56012773513793956 0.93145474791526806 0.25074130296707159 -0.9574308693408965 -1.5799034237861631 0.71947535872459423 0.23862367868423465 2.7287132740020756 -0.63844007253646839 0.39663147926330572 -2.0294094085693355 -1.054078161716461 -1.4436311721801756 -0.68222475051879872 -2.1114803552627559 0.75943568348884594 0.83329862356185924 -1.6327362060546873 0.19620216637849811 -1.4239488244056699 -0.72850388288497914 -0.33718860149383539 -0.73277828097343434 0.34618109464645391 -0.57440119981765736 -1.25513219833374 2.0766216516494755 0.5201290249824525 -0.47939994931221003
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 20 3 4 18 -6 8 15 -7 13 11 -11 -12 14 -4 -3 17 -5 -2 -8 -1 -13 -20 -15 29 -23 -22 -28 -18 -25
right_child=2 7 9 16 5 6 19 -9 -10 10 12 21 -14 23 -16 -17 28 -19 22 -21 26 25 -24 24 -26 -27 27 -29 -30 -31
leaf_value=0.057144949459584385 -0.03790556325170974 -0.14350926574053652 -0.12292232003512711 0.10967637882762155 -0.1266806731815146 -0.099587855580093809 0.067148369511684 0.087406367203518312 0.059398559571151033 0.045932143435390128 -0.068655426758704244 -0.066489396009615134 0.1385015970645179 -0.089533103429877303 0.044615939046671403 -0.0046037156465631515 0.16862102716747385 -0.021983318612291389 -0.073843199422173031 0.17279832841998438 0.11502161694602647 -0.1137630948543438 -0.15702986721494749 -0.15260173492064533 -0.12624491617074698 -0.15986355028908641 0.15320468172320101 0.1290555109510361 0.15301642665963991 -0.16306862178616921
leaf_weight=5.4622537493705741 6.6336526125669506 7.3332201838493356 5.0631619244813919 8.5321358442306536 7.7764407247304908 8.7329181730747241 9.0184726566076261 7.9165684282779685 6.0240017026662818 4.8231356441974667 4.6820600777864456 6.2857051640749004 5.1054009050130826 7.7734497040510169 5.9681497216224653 6.1316226571798333 4.9649696797132474 5.1854799985885611 4.5776880830526414 5.7150826156139365 5.567393645644187 4.5835875719785717 12.156179234385485 4.9520807564258602 8.7389129996299726 8.5643428713083214 34.08418644964695 4.4829242527484885 5.6244907528162003 11.493125855922701
leaf_count=23 27 32 22 36 33 36 37 33 25 20 20 26 21 33 25 25 21 22 20 24 24 20 53 22 38 37 151 20 25 49
internal_value=0 0.089665 -0.0390804 0.00138193 -0.0396183 0.00257956 0.0366646 -0.0181772 -0.0346872 -0.0831249 -0.0499735 -0.0860364 0.0394032 -0.108782 -0.0322809 -0.0802544 0.103658 0.0599069 -0.106916 0.10813 0.136156 -0.118789 -0.134273 -0.134388 -0.148233 -0.143792 0.145935 0.150398 0.160333 -0.159917
internal_weight=0 70.9782 162.975 84.9415 60.6344 37.2669 29.4905 21.3814 14.7569 78.0331 34.0442 24.2568 9.78746 43.9889 11.0313 13.4648 24.3071 13.7176 23.3675 14.7336 49.5968 19.4336 16.7339 32.9576 25.1841 13.1479 44.1345 38.5671 10.5895 16.4452
internal_count=1000 308 692 359 255 155 122 90 61 333 144 103 41 189 47 57 104 58 100 61 218 83 73 142 109 57 195 171 46 71
is_linear=0
shrinkage=0.1
Tree=5
num_leaves=31
num_cat=0
split_feature=8 8 7 4 3 0 3 0 0 0 1 8 0 31 7 31 7 5 20 4 26 0 3 4 30 27 14 26 19 15
split_gain=69.2578 25.9303 30.0573 23.3887 13.5407 9.32488 10.399 9.13313 9.00562 6.25418 10.3027 9.26236 7.74622 6.67162 5.89016 5.66404 5.61037 5.21913 4.97194 4.95097 4.33977 3.18555 4.65588 2.73427 2.19274 1.66334 1.39564 0.567897 0.517343 0.219307
threshold=-1.1515020728111265 0.93145474791526806 0.25074130296707159 0.74304640293121349 1.0758628249168398 0.25519192218780523 -0.60125130414962757 0.23862367868423465 -0.87671247124671925 -0.53437867760658253 0.31563207507133489 2.4709751605987553 1.6993983983993532 0.52829435467720043 -2.1114803552627559 0.46118767559528356 0.72382399439811718 0.68198090791702282 0.46808058023452764 -1.3408321142196653 -0.16108404099941251 -0.43635436892509455 1.2886238098144533 0.52956873178482067 0.11919147521257402 -0.8299041986465453 0.64732486009597789 -0.22411368787288663 -0.77139458060264576 1.1410110592842104
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=3 2 4 21 5 15 18 14 -6 25 11 12 13 -11 -5 26 27 19 24 -4 -13 22 -1 -10 -7 -3 -2 -12 -27 -23
right_child=1 9 17 7 8 6 -8 -9 23 10 16 20 -14 -15 -16 -17 -18 -19 -20 -21 -22 29 -24 -25 -26 28 -28 -29 -30 -31
leaf_value=0.0070112627821202218 -0.16007954657423551 -0.071877512455309531 0.13712212833593604 -0.13228760178207799 -0.062708837026831124 -0.13947779800554305 0.065379170948560009 0.08620552001177291 0.15056030468114676 -0.085946275391551377 -0.10894197577449943 -0.16419502548494241 0.100785898162566 0.062187404902534653 0.013837647351984223 -0.031653448282884168 -0.022768718609493883 0.15139470783869624 0.01825964633699639 0.012918541684253111 -0.040977490231550895 0.1456038744820852 0.13521422858632953 0.049173859296189094 -0.049845080746505024 -0.10785517129231961 -0.097058860958300255 -0.15040904378934017 -0.14782788226782745 0.12135065943483288
leaf_weight=4.8474905937910062 10.51203712821007 5.0448000133037594 5.0108862519264212 5.9062305539846429 5.0222423076629648 6.6929214447736758 9.4710019975900668 6.3724285811185828 5.3085495084524119 7.7890752553939873 4.5042539536953035 4.9416887760162354 7.3390656262636176 4.9869197160005569 5.1759803593158731 7.1344368159770957 5.4641200006008139 10.318273887038229 4.83334857225418 8.9269337207078951 6.7802390903234464 29.530091315507889 6.8155194818973541 5.3315156400203705 4.6086958944797516 4.3848441392183348 5.2785929441452017 12.380144089460371 12.377385124564167 4.2670692056417456
leaf_count=20 46 23 22 27 21 30 39 27 24 33 20 21 30 21 21 31 23 47 20 38 29 138 31 22 20 20 23 56 57 20
internal_value=0 -0.0341579 0.00264952 0.089211 -0.0331843 -0.0592755 -0.0177995 -0.00918719 0.0476615 -0.0769989 -0.0587995 -0.022266 0.018909 -0.0281245 -0.0640394 -0.105602 -0.110844 0.0974831 -0.0666242 0.0575719 -0.0929231 0.126991 0.0819293 0.0997577 -0.102926 -0.12222 -0.139013 -0.139347 -0.137371 0.142542
internal_weight=0 164.442 88.4494 62.9148 64.1933 48.531 25.606 17.4546 15.6623 75.9925 54.1855 31.837 20.1151 12.776 11.0822 22.9251 22.3485 24.2561 16.135 13.9378 11.7219 45.4602 11.663 10.6401 11.3016 21.807 15.7906 16.8844 16.7622 33.7972
internal_count=1000 716 383 284 276 209 109 75 67 333 233 134 84 54 48 100 99 107 70 60 50 209 51 46 50 100 69 76 77 158
is_linear=0
shrinkage=0.1
Tree=6
num_leaves=31
num_cat=0
split_feature=8 4 8 3 10 0 7 9 11 10 2 1 24 4 0 7 0 20 10 1 21 15 10 8 29 18 5 14 5 15
split_gain=58.9011 30.6555 18.6155 29.7394 15.72 12.2131 10.9322 8.16402 7.82123 5.91099 6.5925 6.87362 6.5596 5.69297 5.99512 5.49667 5.24869 5.14153 4.22653 3.54898 3.48699 3.18343 2.78258 2.06416 1.18287 1.16755 1.7522 0.638262 0.249747 0.214235
threshold=-0.82658091187477101 0.45788866281509405 1.1326235532760622 -0.9574308693408965 4.5763163566589364 0.23862367868423465 0.25074130296707159 0.71947535872459423 1.9994604587554934 2.586714625358582 2.9853472709655766 -2.1899367570877071 0.56038278341293346 -0.020042777061462399 -0.641687512397766 -2.1114803552627559 -1.4239488244056699 -0.20484713464975354 -4.1106836795806876 1.0000000180025095e-35 0.81614863872528087 0.11835835874080659 1.1321293115615847 0.38639461994171148 -0.54002109169960011 -0.18906012177467343 -0.80016306042671193 0.9970557987689973 2.0766216516494755 -0.23903460800647733
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 16 3 18 6 15 7 8 21 10 11 -4 22 -9 -15 -3 -1 -14 -2 -7 24 -5 25 29 -20 26 -13 28 -18 -8
right_child=2 5 9 4 -6 19 23 13 -10 -11 -12 12 17 14 -16 -17 27 -19 20 -21 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.034309877636446474 -0.0097003030382814773 -0.13227950378240466 0.0070115516470796412 -0.14955462351394663 -0.12618824401157752 0.13437525768688979 0.12086466180225604 0.15096701267713222 0.053567417576736102 -0.14294136522947901 0.053685659098100386 -0.14752012769801093 0.050369975486841793 -0.055610143137706382 0.085241172809094407 -0.0092659664116992967 0.14259624225412462 -0.084219109749901488 -0.092578399467207584 0.011439473697369737 -0.037576764174953058 -0.033066413876801017 -0.052048141959388908 0.067352653180564659 -0.15357835174809256 -0.15072089254972632 -0.072625898596152136 0.10802476051884853 0.11624473163187055 0.14876358473752652
leaf_weight=6.1406259983777991 4.7501211911439922 7.816032350063324 8.9762760698795301 4.7106884121894819 5.4190714359283438 4.5565567463636381 4.2169402092695245 5.3832936584949485 6.1477786749601355 10.829841911792753 4.5117450356483451 5.794651255011555 5.0614321678876877 5.0904531478881818 7.4363373517990148 6.7862105667591113 29.545384287834167 6.4624808281660062 4.4013430327177092 4.8453724682331085 4.5807319879531851 4.6735247373580933 5.9107572436332694 5.9934198856353751 11.445437714457508 12.035169482231145 6.7774663269519824 8.0390609651803953 4.0950758010148993 7.9253921955823907
leaf_count=27 21 37 39 20 24 20 20 24 26 52 20 27 22 22 31 28 144 29 20 21 20 20 27 26 52 55 30 39 20 37
internal_value=0 0.0743359 -0.0359645 -0.00415128 0.0358321 -0.0178746 0.0528549 0.0189516 -0.0341052 -0.0753595 -0.0621792 -0.0724256 -0.089386 0.0649635 0.0280041 -0.0751105 0.120623 -0.0251061 -0.0946653 0.0710191 -0.114423 -0.0915412 -0.113659 0.115372 -0.136636 -0.128458 -0.107146 0.133339 0.139388 0.139074
internal_weight=0 71.8243 148.534 82.1745 56.9969 24.0042 51.5778 33.4421 15.532 66.3598 55.53 51.0182 42.042 17.9101 12.5268 14.6022 47.8201 11.5239 25.1776 9.40193 20.4275 9.38421 30.518 18.1358 15.8468 24.6073 12.5721 41.6795 33.6405 12.1423
internal_count=1000 336 664 363 250 106 226 143 66 301 249 229 190 77 53 65 230 51 113 41 92 40 139 83 72 112 57 203 164 57
is_linear=0
shrinkage=0.1
Tree=7
num_leaves=31
num_cat=0
split_feature=8 8 7 4 3 3 6 4 9 11 0 11 6 0 9 4 6 1 24 5 9 5 0 0 6 1 2 26 3 13
split_gain=50.3188 18.3556 21.7068 17.5301 9.97278 8.45677 7.25736 9.15673 12.479 7.23407 6.47177 6.01697 5.64528 7.21064 6.00338 15.0436 8.51955 9.95834 5.18458 4.5797 4.36847 3.95235 3.63954 3.02611 2.63279 2.56145 2.12875 2.83894 1.84926 0.264571
threshold=-1.1515020728111265 0.93145474791526806 0.25074130296707159 1.0411096811294558 1.0758628249168398 -0.22925552725791928 2.2315721511840825 -0.16142201423645017 0.31265056133270269 4.0696237087249765 -0.87671247124671925 -0.38942228257656092 0.90315422415733349 -1.3959577679634092 -2.2834039926528926 -2.1853830814361568 -1.5970541238784788 -0.53732573986053456 -0.39458091557025904 0.68198090791702282 -0.85395133495330799 -0.51409405469894398 -1.4239488244056699 1.5622318387031557 0.6653231978416444 -0.35371389985084528 1.3687226772308352 0.053147282451391227 -1.2419523596763609 1.0895674824714663
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=3 2 4 22 6 -5 7 8 11 26 -6 -2 13 -3 -15 -16 25 18 -18 20 -4 -7 -1 -14 -12 -17 27 -9 -24 -30
right_child=1 12 19 5 10 21 -8 9 -10 -11 24 -13 23 14 15 16 17 -19 -20 -21 -22 -23 28 -25 -26 -27 -28 -29 29 -31
leaf_value=0.033096160157524229 -0.10348001834040726 -0.12721967927688255 0.12768464458081583 -0.10930591282215535 -0.052160190085801576 -0.023443534079693604 -0.15190872610015932 -0.027163327387997205 0.13430067521801911 0.036825434455344197 0.13413267755068509 0.044840846020974628 -0.14259706241531853 -0.10746317766558085 0.14272661837861525 -0.15116330182116564 -0.0030146776909804815 -0.079188617224441524 0.1430975565851221 0.13991801900229164 0.0069513267418170434 0.10500525597112079 0.068170243131305597 -0.059636522221130188 0.031304464038983328 -0.05355250294481223 -0.14424512990705093 -0.13303531234074084 0.13590700950010254 0.1077476290773588
leaf_weight=5.0342929661273947 6.7471999526023874 9.7792184948921186 4.5631361901760092 5.640211522579194 4.8819630891084662 4.5014976263046256 6.2405068427324286 5.5568623393774121 6.0684242248535147 4.4246772378683081 5.5661255866289121 4.5996538996696446 12.366793617606165 8.4699546247720701 5.2280920892953828 6.4671130478382111 4.8812817633151981 8.3612256497144788 4.8330061435699472 9.2849135994911176 8.7314402163028735 5.1202752143144581 5.1714498549699774 6.8224933296442023 4.5054202079772949 4.6010079085826856 8.0260421335697156 4.6540065109729767 28.385179132223129 3.7809662073850623
leaf_count=22 30 50 21 27 21 20 29 25 26 20 26 20 60 39 22 30 21 39 20 47 39 22 25 31 20 21 37 22 148 20
internal_value=0 -0.0295646 0.00221407 0.0798125 -0.0286729 -0.0120812 -0.0515964 -0.0359764 0.0185495 -0.0778792 0.0423309 -0.0433555 -0.0666715 -0.0497398 -0.0320539 -0.0134714 -0.0414919 0.000816684 0.0696784 0.0860278 0.048391 0.0449111 0.112912 -0.113102 0.0881334 -0.110587 -0.105709 -0.0754187 0.123674 0.132597
internal_weight=0 155.661 83.8504 57.6339 61.2709 15.262 46.3174 40.0769 17.4153 22.6616 14.9535 11.3469 71.8102 52.6209 42.8417 34.3717 29.1436 18.0755 9.71429 22.5795 13.2946 9.62177 42.3719 19.1893 10.0715 11.0681 18.2369 10.2109 37.3376 32.1661
internal_count=1000 716 383 284 276 69 209 180 76 104 67 50 333 242 192 153 131 80 41 107 60 42 215 91 46 51 84 47 193 168
is_linear=0
shrinkage=0.1
Tree=8
num_leaves=31
num_cat=0
split_feature=8 0 0 8 3 7 7 9 3 3 11 9 6 7 20 17 4 11 24 10 6 17 2 6 1 8 2 3 24 8
split_gain=42.7186 16.1163 16.0962 15.9541 23.7222 11.713 9.51017 8.71349 8.91699 8.71342 6.34335 6.17266 9.25193 9.14627 6.05773 5.03449 4.25775 3.66931 3.38354 4.45809 3.19066 2.92451 2.69306 2.33014 2.20513 0.969835 0.770254 0.575474 0.348763 0.166231
threshold=-1.1515020728111265 -1.608339190483093 -1.8935308456420896 1.4360643625259402 -0.94009315967559803 0.19911569356918338 -1.6642212867736814 0.82368111610412609 0.12777689099311831 -0.19423490762710569 1.4761498570442202 -0.036320179700851433 -1.5970541238784788 -0.59667176008224476 0.47881735861301428 -0.38247415423393244 -1.8773684501647947 1.784462094306946 0.42835316061973577 0.46801452338695532 0.77008655667305004 0.25972323119640356 1.5823640227317812 -1.3626608252525327 0.26765033602714544 -0.28038287162780756 2.6751513481140141 -0.27606296539306635 -0.43549998104572291 0.13030955195426944
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=2 24 -1 4 14 7 9 8 -6 -4 -10 18 -13 -14 16 -9 -3 -11 23 -20 28 -17 -16 -5 27 -18 -8 -2 -7 -30
right_child=1 3 6 11 5 20 26 15 10 17 -12 12 13 -15 22 21 25 -19 19 -21 -22 -23 -24 -25 -26 -27 -28 -29 29 -31
leaf_value=-0.073968990352853564 -0.093739022207804903 -0.010974665416263824 -0.055367137139929495 -0.061695757240100818 -0.10780484118785466 0.11278655725669283 0.13459008019950219 -0.022574204767896192 -0.05746098793325604 0.15056556971718599 0.086481534268267032 -0.11451920845370428 0.13940290555941964 -0.04354142169523384 -0.050608647379268561 0.14956166878271104 -0.090207742662018364 0.031335665898348429 0.026265675534095756 -0.11806096099369523 0.045777414493210683 0.051425721454271295 0.057394410072918282 -0.14058046548058656 -0.057914127689205286 -0.14248115906458689 0.095083757265389229 -0.13808940312949178 0.13575844352450489 0.16175852764178442
leaf_weight=6.3895061016082755 3.9089904278516796 4.609561592340472 6.6107281893491772 5.289188146591191 8.2041784971952456 3.9828253388404855 25.74672427773476 4.2842177301645306 5.2673320323228818 5.2250744849443418 7.3108304589986801 6.149515017867091 4.9554191827774057 6.0928271710872624 4.5951239019632322 6.4620812237262708 5.999110251665118 5.1010230630636215 4.5680959522724134 4.0268123298883438 5.1778778582811347 5.7286854386329651 4.6400246173143387 12.821912527084351 6.5930055975913993 8.6912243962287885 6.1054493337869635 11.631036579608915 5.2937370836734772 4.5921443700790405
leaf_count=29 20 20 32 24 38 21 139 20 23 27 32 30 21 29 21 29 30 25 20 20 25 25 20 64 32 43 32 60 28 21
internal_value=0 -0.0275489 0.0752775 -0.0139976 0.0113262 0.0489705 0.0948231 0.0163593 -0.0266988 0.034277 0.026203 -0.0629327 -0.0162072 0.0385138 -0.0629518 0.0706749 -0.0948238 0.0916668 -0.0930224 -0.041353 0.112762 0.103446 0.00365543 -0.117543 -0.106374 -0.121134 0.127017 -0.126933 0.13777 0.147836
internal_weight=0 150.876 55.1785 128.743 84.839 56.3039 48.789 37.2573 20.7823 16.9368 12.5782 43.9038 17.1978 11.0482 28.535 16.475 19.2999 10.3261 26.706 8.59491 19.0466 12.1908 9.23515 18.1111 22.133 14.6903 31.8522 15.54 13.8687 9.88588
internal_count=1000 716 284 604 396 262 255 167 93 84 55 208 80 50 134 74 93 52 128 40 95 54 41 88 112 73 171 80 70 49
is_linear=0
shrinkage=0.1
Tree=9
num_leaves=31
num_cat=0
split_feature=8 7 4 0 9 6 9 0 1 8 0 4 7 8 11 21 16 1 6 9 3 9 15 10 4 14 5 25 6 0
split_gain=36.8622 31.2409 16.111 9.07769 8.31071 8.24479 7.71666 7.25363 10.0546 6.4032 6.29972 5.97859 5.84786 12.5459 6.06547 5.4406 3.9077 3.45596 5.41924 6.19242 2.87959 2.90049 2.86791 1.73676 2.49868 1.59774 1.33031 0.874248 0.790689 0.175107
threshold=0.36960989236831671 -0.17649734020233152 -0.16142201423645017 -1.2200167775154112 1.3687919378280642 -1.5258474349975584 -0.072343081235885606 -1.5481609106063841 1.0527601242065432 -1.8228302001953123 0.37160789966583258 0.90728291869163524 0.72382399439811718 2.3894670009613042 4.8875966072082528 0.071496769785881056 0.039843803271651275 -0.31010645627975458 -0.80244350433349598 -0.89385390281677235 -0.9943212866783141 2.1439498662948613 -0.94553837180137623 0.69830441474914562 -1.6327362060546873 -0.2194641828536987 0.21785226464271548 -0.42061229050159449 1.7042843699455263 -0.90527260303497303
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 2 6 26 12 -6 10 25 9 -9 -1 -10 14 -14 17 -12 -11 18 -5 -20 -3 28 -19 -8 -25 -4 -2 -24 29 -22
right_child=3 20 7 4 5 -7 23 8 11 16 15 -13 13 -15 -16 -17 -18 22 19 -21 21 -23 27 24 -26 -27 -28 -29 -30 -31
leaf_value=-0.083349666608681802 -0.13444854567981296 0.044240472395180475 -0.065774594756563692 -0.12233548598536155 -0.075813236943054751 0.095741552185865803 0.15233993833638618 0.11836466835138422 0.0059343848649581346 0.046254312557815293 0.14507854880819007 -0.14137463793728397 0.095002042846111467 -0.092614842450050403 0.033907414135622281 -0.015172342898730668 -0.066112427770854507 -0.039907196829146389 -0.089519974184037607 0.063065290316737771 0.15349046668481173 0.0337772490839287 -0.09296581812600524 0.13769911844947239 0.02822103262581176 -0.14175741376787843 -0.066525178316831537 -0.1405775898675356 0.096860540531114814 0.13104059275959637
leaf_weight=5.2529212534427669 11.769869163632391 6.4662766158580771 4.6920508593320847 6.0131878256797817 4.161924496293067 8.5695785880088788 5.4692756086587897 5.6498281657695797 4.313169986009596 6.1698920875787753 3.7972993701696378 7.6270183473825455 7.534880578517912 6.7634193301200867 5.2381171733140937 4.7922911494970322 6.2097992151975632 5.0847383290529207 4.7392872869968423 6.0614190250635049 4.1650314927101126 3.772762492299079 5.7031664848327628 3.9468927383422843 4.4187788218259803 6.7467360794544202 3.8190941661596289 11.911302521824853 7.0482614487409583 20.952388301491741
leaf_count=24 66 32 23 28 21 39 27 27 22 28 20 36 34 32 24 25 28 24 23 26 20 19 30 22 24 35 20 62 39 120
internal_value=0 0.0381085 -0.00336345 -0.0486406 -0.0336192 0.0396603 0.0557039 -0.0428441 -0.0169869 0.0301494 0.00291617 -0.088162 -0.0494188 0.00625501 -0.0672069 0.0556716 -0.0101102 -0.0806113 -0.0462488 -0.00388824 0.105675 0.116728 -0.106064 0.10852 0.0798724 -0.11059 -0.117808 -0.125162 0.126458 0.134763
internal_weight=0 111.491 69.086 87.37 71.781 12.7315 27.6775 41.4085 29.9697 18.0295 13.8425 11.9402 59.0495 14.2983 44.7512 8.58959 12.3797 39.5131 16.8139 10.8007 42.4047 35.9384 22.6992 13.8349 8.36567 11.4388 15.589 17.6145 32.1657 25.1174
internal_count=1000 571 341 429 343 60 142 199 141 83 69 58 283 66 217 45 56 193 77 49 230 198 116 73 46 58 86 92 179 140
is_linear=0
shrinkage=0.1
Tree=10
num_leaves=31
num_cat=0
split_feature=8 4 6 8 7 11 7 4 1 4 5 1 5 16 0 1 3 0 7 19 27 30 20 14 6 1 19 5 3 8
split_gain=32.357 20.7087 10.58 12.6193 12.4144 9.07447 7.99128 7.40886 6.17225 6.16662 9.60891 5.6191 5.02313 4.24238 4.11905 5.8183 4.11393 3.78869 3.59757 3.33792 2.01766 0.898773 0.790267 0.768399 0.762128 0.697344 0.652533 0.443194 0.0899022 0.0865904
threshold=-0.82658091187477101 0.45788866281509405 1.7042843699455263 1.4360643625259402 0.25074130296707159 0.10654639080166818 -1.1648204326629636 -2.4063184261322017 0.47342592477798467 -0.38929480314254755 -0.67728805541992176 -1.6409657001495359 -0.4936488270759582 -0.039748009294271462 -0.32855796813964838 -1.6409657001495359 -0.63477313518524159 -1.4239488244056699 -1.5511026382446287 0.66233554482460033 0.25801476836204534 0.38249950110912329 -0.41391812264919275 1.3252758979797366 0.03198367357254029 2.8490761518478398 0.55610954761505138 -0.17103919386863706 -0.36424401402473444 0.19538971781730655
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 17 3 4 5 7 8 -2 -3 -7 18 -9 -6 16 26 -16 -4 -1 -11 20 24 -15 -14 25 -17 28 -5 -13 -19 -24
right_child=2 6 13 14 12 9 -8 11 -10 10 -12 27 22 21 15 19 -18 23 -20 -21 -22 -23 29 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.028856137766449366 0.052148750896673292 0.00020451385175358395 -0.11207912302411645 -0.12806508406133549 0.014099436150263159 0.10388266339805939 0.071340593984151643 -0.018206375841594395 -0.12899631887661508 -0.021304282791095776 0.078763791683420431 -0.099161436135509526 0.084651427918297009 -0.16170140679383413 0.034810716282655536 -0.13774789851948371 0.014831875385370103 0.13803743944930577 -0.14490741155922102 -0.003995427373596522 -0.046746419776609559 -0.1031293010098896 0.12984961849455229 0.070282739518784373 -0.079809940926267475 0.085264856288683771 -0.075272205560486291 -0.1400944499001473 0.12466319133923803 0.15116507624420419
leaf_weight=5.7213220149278632 4.9131297618150738 8.0806474834680575 4.9776752293109894 7.5145658105611783 7.0344296097755459 7.4244205206632641 6.8876379430294028 7.7304381877183959 6.8167511671781531 5.9360385388135946 7.7087390273809406 3.8706749975681349 3.8936640173196801 6.2017563730478269 7.7555155456066158 6.4642212390899711 5.2462612986564654 6.8678615391254452 3.9031166285276413 6.431323066353797 6.8009473830461493 4.5358942896127701 4.000351965427396 3.1882668137550345 3.4995084255933762 4.4264312535524359 3.4008550643920898 8.3541819602250964 18.742216736078266 3.6399082541465759
leaf_count=27 22 39 25 45 35 35 33 38 34 28 36 21 21 30 35 34 26 37 20 30 34 24 24 19 20 27 20 43 120 18
internal_value=0 0.0603441 -0.0279363 -0.0155573 0.0109067 -0.0150949 -0.017733 -0.0578535 -0.0589152 0.0274858 -0.00483728 -0.0849368 0.0806995 -0.0930606 -0.0587985 -0.0401715 -0.0469566 0.104018 -0.0703367 -0.0652416 -0.0887371 -0.136959 0.121318 0.11696 -0.117399 0.121915 -0.111617 -0.127134 0.12825 0.140005
internal_weight=0 60.7311 131.238 110.276 68.4091 49.8407 21.785 24.8684 14.8974 24.9723 17.5479 19.9553 18.5684 20.9616 41.8669 30.9515 10.2239 38.9461 9.83916 23.196 16.7647 10.7377 11.5339 33.2248 9.96373 30.0365 10.9154 12.2249 25.6101 7.64026
internal_count=1000 336 664 559 341 243 106 124 73 119 84 102 98 105 218 153 51 230 48 118 88 54 63 203 54 184 65 64 157 42
is_linear=0
shrinkage=0.1
Tree=11
num_leaves=31
num_cat=0
split_feature=8 0 7 3 6 8 3 7 4 4 9 5 9 6 7 5 9 18 24 19 6 9 20 1 6 6 10 8 4 9
split_gain=27.8512 11.8014 11.7649 10.3329 10.1752 13.0742 13.6968 8.0119 6.79693 6.69961 5.898 5.78428 4.60364 7.19059 6.21579 4.57637 4.11375 3.97437 3.95228 3.88977 3.54683 3.37432 2.53909 2.05013 1.77695 1.20771 0.587259 0.574022 0.255189 0.15827
threshold=-1.1515020728111265 -1.608339190483093 -1.6642212867736814 -0.19423490762710569 2.2315721511840825 1.4360643625259402 -0.94009315967559803 -0.13924100995063779 -0.1948781609535217 -1.5903732776641843 1.2183623313903811 -0.69947561621665943 -0.036320179700851433 -1.5970541238784788 -0.59667176008224476 -0.42181968688964838 -1.6292568445205686 0.018532840535044674 0.71080669760704052 -0.40020120143890375 -1.3626608252525327 2.3375082015991215 0.5318211317062379 0.26765033602714544 0.32099404931068426 2.1336631774902348 -0.028142143040895459 -0.28038287162780756 0.41396835446357733 -1.502330422401428
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=2 23 3 11 5 6 9 8 15 -3 17 -1 18 -14 -15 -8 -6 -10 20 -5 -7 25 27 26 29 28 -2 -11 -4 -9
right_child=1 4 21 19 16 12 7 24 10 22 -12 -13 13 14 -16 -17 -18 -19 -20 -21 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.0030969595351530935 -0.081380366702221985 0.035807401456367856 0.12449848152453878 -0.0053320943691444024 -0.016565429212879616 -0.031408581705343219 0.13459638232855314 0.11183553152570365 -0.13295369777172947 -0.086205728779757393 0.045834693166422913 -0.16212761804617648 -0.10105939133277092 0.12342028637284937 -0.031627410466516091 0.0077237310870855412 -0.14411397669955459 -0.0066618430458213812 0.0012236113537026634 0.1196269762391996 -0.13091004106879203 0.010412477543348248 -0.02424265896231316 -0.046812524935748522 0.064872910558552307 0.079518653974368594 -0.12953723496743538 -0.13660285351618365 0.1547327341945825 0.13813668466158369
leaf_weight=4.316520556807518 3.3995198309421566 7.1535336971282986 16.165391221642505 4.1005556732416153 3.8101630955934525 5.1121072471141842 6.8718056380748749 3.1714788824319848 4.8000036329031008 4.0793030112981796 8.5335142910480482 4.1616955548524874 5.4241134673357001 4.8641585111618033 5.51972672343254 4.8493687957525271 7.5180331468582136 5.1818968802690506 4.8408790975809088 6.3466993719339353 11.972646012902262 3.3262214064598075 4.787720948457717 5.9887080639600745 6.3763602375984183 6.3494299352169028 9.9265171587467176 5.0676726251840574 3.3744001984596244 8.2128507345914841
leaf_count=22 20 32 108 21 20 25 34 21 23 23 38 22 30 21 28 24 39 24 23 37 69 18 22 32 34 39 60 28 17 46
internal_value=0 -0.0229445 0.065438 0.0040165 -0.0110975 -0.00154066 0.0243149 0.0538289 0.022516 -0.0428597 -0.0152072 -0.0780067 -0.0488789 -0.00774269 0.0410021 0.0821057 -0.101214 -0.0673919 -0.0785373 0.0705804 -0.101137 0.105226 -0.0832444 -0.0954118 0.107137 0.117408 -0.117252 -0.114127 0.12972 0.13081
internal_weight=0 137.462 48.1409 18.9255 118.147 106.819 69.0855 47.9973 30.2366 21.0882 18.5154 8.47822 37.7336 15.808 10.3839 11.7212 11.3282 9.9819 21.9256 10.4473 17.0848 29.2154 13.9347 19.3147 17.7607 25.8892 13.326 9.14698 19.5398 11.3843
internal_count=1000 716 284 102 604 545 349 244 143 105 85 44 196 79 49 58 59 47 117 58 94 182 73 112 101 164 80 51 125 67
is_linear=0
shrinkage=0.1
Tree=12
num_leaves=31
num_cat=0
split_feature=8 7 3 0 0 1 22 1 4 2 11 8 14 8 14 20 6 4 16 26 8 9 16 26 6 8 26 24 20 29
split_gain=24.4224 21.9039 11.1647 9.23084 7.17254 7.56577 7.24002 6.99716 8.01382 6.65717 4.69433 4.60934 4.46166 3.69161 3.7488 3.47332 3.0632 2.7334 7.6099 2.66246 2.50286 2.25209 1.91924 1.65004 2.57066 0.897843 0.648219 0.561147 0.489143 0.487628
threshold=0.36960989236831671 -0.17649734020233152 -1.5538635253906248 -1.0424582958221433 -0.53437867760658253 1.2998940348625185 0.81031483411788952 0.82854908704757702 0.74304640293121349 1.3687226772308352 3.7105647325515752 -1.4973685741424558 1.1271307468414309 1.5303894281387331 -0.10719700530171393 0.14777838438749316 -0.65812426805496205 0.45788866281509405 -0.053894041106104844 -0.3003257811069488 -0.043257951736450188 2.2335890531539921 -0.52745679020881642 0.3114320188760758 -1.2840110063552854 -0.52487343549728382 0.22822204977273944 -0.58174118399620045 -0.47140231728553766 -0.4948394894599914
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 2 19 9 23 6 10 17 11 -4 12 -9 13 14 -6 -8 -7 28 -19 -1 21 25 -15 27 -25 -3 -21 -2 -5 -18
right_child=4 20 3 7 5 16 15 8 -10 -11 -12 -13 -14 22 -16 -17 29 18 -20 26 -22 -23 -24 24 -26 -27 -28 -29 -30 -31
leaf_value=-0.026662588811133415 -0.085080521902390072 0.12441039896962042 0.016453353521028106 0.089361142741479882 0.039096426470727606 -0.022674501965035437 0.017257765887608769 0.12350717142456782 -0.11041172444604413 -0.12000900705692691 0.054566465319364411 -0.018888302373026879 0.054664854906428285 -0.035459024644611127 -0.078827301829477739 0.13470585986007796 -0.090218874193332771 0.12875527897956979 -0.043788078804843068 -0.083498111478987988 0.034715684262131348 0.025237881089246173 -0.11895498726965976 -0.12315819830716222 -0.014830686411872421 0.079080197549749501 -0.13973221341475911 -0.13134720717763512 0.13476125425539145 -0.13795672958470559
leaf_weight=5.331623136997222 3.9050936549901962 21.984253019094471 7.5827694684267026 3.5700577050447411 5.8253472298383713 4.8452204614877692 5.4925420209765417 4.2950438633561125 4.8996021300554267 6.7635987848043442 6.0792963802814475 4.8292034789919844 4.0564924776554099 3.911944553256034 5.0180152207613045 4.6494683474302292 3.3443343117833129 5.4082370549440384 4.8469817638397199 3.458766818046568 5.8154154866933814 3.081442102789878 9.2920361757278425 3.7249053046107292 5.3183401972055417 5.4532888159155872 5.0322006046772003 7.9748657494783419 7.0783106684684824 5.9406574815511695
leaf_count=29 26 147 38 22 28 25 27 29 23 36 28 25 21 20 26 22 20 25 21 21 32 17 51 25 29 34 29 49 43 32
internal_value=0 0.0329899 -0.00262711 0.0196525 -0.0413937 -0.0234097 -0.00309714 0.0473921 -0.00725338 -0.0478819 -0.0251112 0.0481413 -0.0423468 -0.0587114 -0.0154755 0.0711003 -0.0871281 0.0840528 0.0472051 -0.082048 0.0948405 0.106297 -0.0942176 -0.0916375 -0.0594507 0.115401 -0.116825 -0.116139 0.11954 -0.120762
internal_weight=0 99.4308 63.0964 49.2738 79.3786 58.4554 44.3251 34.9274 14.0238 14.3464 34.1831 9.12425 28.1038 24.0473 10.8434 10.142 14.1302 20.9036 10.2552 13.8226 36.3344 30.519 13.204 20.9232 9.04325 27.4375 8.49097 11.88 10.6484 9.28499
internal_count=1000 571 341 262 429 300 223 188 77 74 174 54 146 125 54 49 77 111 46 79 230 198 71 129 54 181 50 75 65 52
is_linear=0
shrinkage=0.1
Tree=13
num_leaves=31
num_cat=0
split_feature=8 4 5 11 7 9 0 0 6 6 7 6 9 6 23 8 3 9 11 0 3 12 7 1 24 16 16 30 18 10
split_gain=20.7184 15.9591 8.9729 11.7562 10.8995 11.2904 6.48128 6.45537 5.95731 8.29799 8.37787 4.85174 4.67074 7.48223 4.62533 4.2342 7.44244 3.61216 3.51804 3.27023 5.92708 2.85761 2.66045 2.54425 1.74062 1.4283 1.35726 0.597551 0.383279 0.324064
threshold=-0.82658091187477101 0.45788866281509405 -0.61190077662467945 0.10654639080166818 0.72382399439811718 -0.25800773501396174 -0.37390863895416254 0.23862367868423465 -1.7323389053344724 1.7042843699455263 -1.8709572553634641 -1.7323389053344724 -1.9994176626205442 0.32099404931068426 0.29328140616416937 1.0175021886825564 -0.22925552725791928 0.82368111610412609 1.9994604587554934 -0.43635436892509455 0.37997007369995123 -0.73277828097343434 -2.577299833297729 1.0000000180025095e-35 -0.0057606240734457961 -0.1045374535024166 0.3629890382289887 -0.54829272627830494 1.2640730142593386 0.80241575837135326
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 19 8 4 5 27 24 22 -2 10 17 -6 -5 14 -14 16 -12 -10 -18 20 -1 -17 -3 -9 -7 -11 29 -4 -21 -23
right_child=2 7 3 12 11 6 -8 23 9 25 15 -13 13 -15 -16 21 18 -19 -20 28 -22 26 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.041581438079044908 -0.12187154513500512 -0.12765314804558797 -0.084380549491176182 0.13708726113588585 0.14295876155175746 -0.11361506854730737 0.075798404785022214 0.11156945322599805 0.0029196891760688232 -0.07268489671271218 -0.084493238535091567 -0.0028263398823203593 0.15097929321636777 -0.073559161079991123 0.010922775578712714 -0.0014109302589292395 -0.0034913679440021677 0.12017637119170169 0.13043348315566292 0.12054567370493506 0.11361963427846795 -0.13685947817641256 -0.029536434792624668 -0.0019388729551211297 -0.012712958572466094 -0.14028482038906404 -0.049769561201914669 -0.13190496751411904 0.078419196601304836 -0.095151019049661195
leaf_weight=4.2814871072769147 9.5789865478873235 4.2692422717809686 3.4879172518849364 6.4339557141065624 4.3723699077963829 3.3538016378879547 6.7768559455871582 3.7617974132299414 5.1530557125806755 6.1080239489674568 6.0156893879175186 4.7767762690782529 4.9436333179473859 4.4105255454778662 4.5082527995109558 3.8401504009962038 3.9300314784050041 5.3597636967897415 3.9158187955617905 20.329771213233474 5.7859806939959526 5.5289824232459086 7.8358571231365213 4.1567702889442444 3.4873388037085533 6.4010387361049634 3.6616331860423079 10.95682118088007 2.416472226381301 2.8094499707221985
leaf_count=21 64 27 20 33 26 21 34 20 24 33 33 26 25 23 23 20 20 26 20 153 37 32 38 21 19 36 22 65 19 19
internal_value=0 0.0521364 -0.0230329 0.00545143 -0.0279402 -0.0588422 0.00648393 -0.0182176 -0.0493253 -0.0361449 -0.0140188 0.0668442 0.0666724 0.0339908 0.0841767 -0.0411735 -0.00081177 0.0627008 0.0633498 0.095068 0.0476158 -0.0764934 -0.0641404 0.0519844 -0.0621792 -0.107277 -0.100521 -0.120429 0.11607 -0.122807
internal_weight=0 52.8374 119.811 57.5082 37.2119 28.0627 13.618 20.0237 62.3026 52.7236 40.2146 9.14915 20.2964 13.8624 9.45189 29.7018 13.8615 10.5128 7.84585 32.8137 10.0675 15.8402 12.1051 7.91857 6.84114 12.5091 12.0001 14.4447 22.7462 8.33843
internal_count=1000 336 664 315 211 159 74 106 349 285 216 52 104 71 48 166 73 50 40 230 58 93 65 41 40 69 73 85 172 51
is_linear=0
shrinkage=0.1
Tree=14
num_leaves=31
num_cat=0
split_feature=8 0 0 6 8 7 13 3 7 3 4 13 9 0 10 2 19 4 9 27 13 11 6 29 1 18 2 22 8 4
split_gain=17.9671 9.16898 8.56836 7.70969 8.37988 10.1703 6.81208 7.43396 5.97358 5.45081 4.99929 5.38085 4.32265 4.11448 3.7356 3.74934 4.33141 3.68053 3.42629 3.26615 2.90249 2.69301 1.98768 2.52063 1.9156 1.52486 1.01647 0.611931 0.0996783 0.0453145
threshold=-1.1515020728111265 -1.8935308456420896 -1.608339190483093 2.2315721511840825 1.4360643625259402 0.25074130296707159 1.279394865036011 0.12777689099311831 -1.6642212867736814 -0.19423490762710569 -0.44345408678054804 -0.18393839150667188 1.2301043272018435 3.2111262083053593 2.1139192581176762 1.7050061225891116 0.82704880833625805 -0.14276775717735288 -1.6292568445205686 -0.25377622246742243 0.45762664079666143 1.784462094306946 -0.47526878118515009 -0.092528741806745515 0.26765033602714544 -0.47756856679916376 2.6751513481140141 0.82855579257011425 0.60461205244064342 0.67168915271759044
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 -1 24 4 5 6 7 12 9 -3 -9 -12 17 14 15 16 20 19 -5 -4 25 -11 28 -24 27 -6 29 -2 -7 -10
right_child=2 8 3 18 13 22 -8 10 26 21 11 -13 -14 -15 -16 -17 -18 -19 -20 -21 -22 -23 23 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.062317036898849425 -0.1222942204368439 -0.049998336963566335 -0.10624712149509202 -0.012313626390081225 -0.045388493115543639 0.11935719169840049 0.092414640766011691 0.10294862058384169 0.11709988440507867 0.13696068398766709 0.066813352443162971 -0.068227662510127435 0.022879627974816675 -0.13031620175804573 -0.12470326986553733 0.052175747869016145 0.054379276249613084 -0.1311251026306097 -0.13434906866198751 0.012162264979624153 -0.0082949931329668845 0.022011053228962577 0.11696507833877162 0.010233037681507226 -0.037121496187472985 -0.12870841375727582 0.065739302696608465 -0.067377917085303066 0.1432588172532189 0.13137200798807877
leaf_weight=5.5635349303483954 8.5689424127340335 5.7589531019330016 4.0545059964060792 3.5351394712924957 3.6320753097534224 4.1312959045171764 6.596461407840251 6.2510937973856953 14.757522217929369 3.9098580852150899 5.3645386248826989 6.557468339800832 5.3809577599167815 4.392749771475791 3.437015324831008 6.483071655035018 4.2301486432552329 7.71580243110657 6.5883191898465139 5.4753524512052545 6.4579198956489554 4.2572583630681038 4.0531137734651574 4.8729184195399258 5.4697172939777365 5.5572983175516173 4.4609812200069419 2.6586362048983565 3.0204468220472336 2.6195266693830481
leaf_count=29 62 32 21 20 20 32 34 34 124 27 27 31 28 26 24 34 21 46 39 27 36 25 25 28 32 35 32 18 16 15
internal_value=0 0.0571061 -0.0189912 -0.0086653 -5.47552e-05 0.0214436 -0.00187008 -0.017114 0.0756838 0.0245055 0.0305156 -0.00746333 -0.0553688 -0.0399662 -0.0266468 -0.0138617 -0.0354 -0.0797837 -0.091734 -0.0382154 -0.0596713 0.0770411 0.0901706 0.0586977 -0.0856492 -0.0957764 0.10832 -0.10929 0.129452 0.119251
internal_weight=0 41.3276 124.485 107.788 97.6642 63.474 47.3962 40.7997 35.7641 13.9261 18.1731 11.922 22.6266 34.1903 29.7975 26.3605 19.8774 17.2457 10.1235 9.52986 15.6473 8.16712 16.0778 8.92603 16.6973 9.18937 21.838 11.2276 7.15174 17.377
internal_count=1000 284 716 604 545 349 248 214 255 84 92 58 122 196 170 146 112 94 59 48 91 52 101 53 112 55 171 80 48 139
is_linear=0
shrinkage=0.1
Tree=15
num_leaves=31
num_cat=0
split_feature=8 7 3 4 0 1 11 22 9 11 0 9 4 26 9 1 7 24 5 10 9 3 20 26 10 27 13 17 27 19
split_gain=16.1553 15.9235 7.96072 6.10052 5.42614 5.85235 5.82598 5.54842 5.19066 4.70367 4.30884 4.05866 10.1016 4.36246 3.61284 3.68509 3.45621 3.40532 3.22119 2.90172 2.82607 3.26318 2.63031 2.53625 1.6996 1.54388 1.38882 1.10519 0.850554 1.2471
threshold=0.36960989236831671 -0.17649734020233152 -1.5538635253906248 -0.14276775717735288 -0.53437867760658253 1.2998940348625185 4.6643924713134775 0.81031483411788952 -0.92759951949119557 -0.20719208568334577 -1.5481609106063841 -0.89385390281677235 -1.803136885166168 -0.16108404099941251 0.46225464344024664 0.082951605319976821 -0.47129309177398676 0.20110688358545306 -0.55146434903144825 0.23376951366662982 2.3375082015991215 -0.9943212866783141 0.68299338221549999 -0.3003257811069488 2.6175767183303837 -0.8299041986465453 -0.32162345945835108 0.68529933691024791 0.16166735440492633 0.16267675906419757
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 2 23 8 25 6 7 11 9 -4 24 17 -13 -14 19 26 -15 -6 -9 -12 21 -3 -7 -1 -5 -2 -16 -10 -27 -30
right_child=4 20 3 10 5 22 -8 18 27 -11 14 12 13 16 15 -17 -18 -19 -20 -21 -22 -23 -24 -25 -26 28 -28 -29 29 -31
leaf_value=-0.017921200632328549 -0.030421784916537149 0.023141025200562673 -0.085088011535358332 -0.032291238208619087 -0.13219098662783732 -0.10770417663172192 0.10336131099333568 -0.0079672011276724387 0.12548505365067822 0.072595418835561892 -0.10318356901156522 0.13475925472505476 -0.12643218469757933 0.048934872550518399 0.12699247279307521 -0.0328543014054875 -0.081474509456712793 -0.022677327802597687 0.11742084285490209 0.018735611213065845 -0.010416492114128366 0.11165296104149558 -0.0059230153358106235 -0.11048056505166702 -0.13184373339885389 -0.12454586480960086 0.041991167401142254 0.051525484340562783 -0.036024721055282774 -0.12028619309084319
leaf_weight=4.9122406765818596 4.3158210068941143 5.0767707973718705 3.5756715461611748 3.438605003058913 6.1498211622238141 9.0219812691211683 4.7281536161899558 3.90865807980299 7.913008861243723 4.016998082399371 3.9165716916322735 3.9833168610930434 4.5328212827444068 4.9423304423689816 3.9686836302280444 3.9640849754214287 3.4515391364693722 5.2746866270899773 4.3058421462774259 3.8920623734593391 2.7507096603512755 23.197395026683807 3.5334853678941718 7.4505607262253744 3.4211119264364243 6.6859009191393834 3.7276054322719538 2.7132240533828735 3.8173650428652763 3.2535320594906807
leaf_count=29 29 30 19 20 38 58 23 20 57 25 20 19 26 26 20 20 20 29 22 20 16 184 19 50 22 48 20 19 26 26
internal_value=0 0.0287124 -0.00270696 0.0169958 -0.0351932 -0.0192765 -0.00109188 -0.0146044 0.0614819 -0.00166359 -0.0137877 -0.0355832 -0.00447462 -0.0473791 0.0102253 0.0454773 -0.00468915 -0.0816286 0.0577582 -0.0424153 0.0863464 0.0957602 -0.0790599 -0.073703 -0.0819406 -0.082604 0.0858231 0.106601 -0.0989748 -0.074796
internal_weight=0 87.9353 56.9104 44.5476 71.9053 53.8326 41.2772 36.549 18.2189 7.59267 26.3287 28.3345 16.91 12.9267 19.469 11.6604 8.39387 11.4245 8.2145 7.80863 31.0249 28.2742 12.5555 12.3628 6.85972 18.0726 7.69629 10.6262 13.7568 7.0709
internal_count=1000 571 341 262 429 300 223 200 120 44 142 158 91 72 100 60 46 67 42 40 230 214 77 79 42 129 40 76 100 52
is_linear=0
shrinkage=0.1
Tree=16
num_leaves=31
num_cat=0
split_feature=8 4 0 6 8 7 1 6 6 10 26 24 12 0 9 20 9 4 0 23 7 5 9 16 5 20 29 25 26 18
split_gain=14.1015 10.3248 6.39172 6.31241 6.30113 9.12662 6.59843 6.59425 5.28462 5.07677 4.04167 3.97092 3.95659 3.76128 3.70751 3.59036 3.82919 3.56359 2.70691 5.66192 2.68585 2.58586 2.29234 2.28323 2.10929 2.09127 1.77215 1.10719 0.700639 0.409233
threshold=-0.94087243080139149 0.56012773513793956 -1.608339190483093 2.2315721511840825 2.4709751605987553 0.34028041362762457 -0.93864721059799183 -1.5804541110992429 -0.47526878118515009 -0.43754623830318445 -0.19718474894762036 -0.30356071889400477 0.8265323042869569 0.50370800495147716 0.023247927427291874 -0.63763517141342152 1.1047214865684511 -1.0175390839576719 -0.43635436892509455 0.3062517791986466 -2.1114803552627559 -0.63844007253646839 -1.4426879882812498 0.42295156419277197 0.33765733242034918 0.5318211317062379 0.11635308712720872 0.31395921111106878 -0.57071822881698597 1.2640730142593386
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 18 24 4 5 6 7 -4 -7 17 -6 -10 14 20 23 -9 -17 26 19 -1 -3 -12 -5 -11 28 -19 -8 -13 -2 -20
right_child=2 13 3 22 10 8 9 15 11 12 21 27 -14 -15 -16 16 -18 25 29 -21 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.11248555960632918 -0.06202729152571395 -0.099176451679655789 -0.09408185900577494 -0.026418761149610389 -0.12651397308212353 0.13609837290494933 0.037740838078345627 0.14080857837512753 -0.054730451330054899 0.029378858173443757 -0.070632694268917273 0.036316318842269026 -0.08356604675359719 0.056251505786204664 0.084633363765536027 -0.023580032570026083 0.098758303528166946 -0.12149706301869745 0.11576384573641794 -0.056101896226298686 -6.375985892882826e-05 0.03990189846737522 -0.13009256608119621 -0.093362396513097912 -0.017517838459535178 -0.039344351637545523 -0.055748295233224027 0.12145427674636267 -0.12031885041448503 0.06790100067413378
leaf_weight=4.6799102425575274 2.830914929509162 5.1513856276869783 3.0079263895750072 3.5046851336956024 5.8060804829001418 7.5968605577945736 3.8810633495449984 4.678051285445691 3.8175725042819986 3.6579803377390068 4.4655471667647353 3.355530679225919 3.617363102734088 5.2631835266947737 6.9655478894710541 5.3353114351630104 4.9157704785466194 9.3756184056401199 15.748438000679018 3.4685920476913452 5.8267735689878473 4.0233053490519515 5.4482695087790471 2.5876270607113838 3.8077028542757025 4.6282024532556543 4.2456135004758835 2.80379518866539 7.5912977904081336 2.0149379000067702
leaf_count=31 21 36 19 21 37 52 20 24 22 21 30 21 22 29 36 27 28 58 147 22 28 22 35 19 24 24 24 19 62 19
internal_value=0 0.0492617 -0.0185981 -0.00947975 -0.00140789 0.0102653 -0.00919081 0.0409981 0.0732556 -0.0322982 -0.0622197 0.0254043 0.0090982 -0.0132505 0.0344707 0.068214 0.0350857 -0.0637769 0.0884438 0.0407225 -0.0465713 -0.0182446 -0.0895089 -0.0214743 -0.0812143 -0.0943459 -0.0111006 0.0750721 -0.104486 0.110335
internal_weight=0 42.1532 111.948 97.7177 88.7648 74.4698 56.8961 17.9371 17.5738 38.959 14.2949 9.9769 16.8285 16.2413 13.2112 14.9291 10.2511 22.1305 25.9119 8.1485 10.9782 8.48885 8.95295 6.24561 14.2299 14.0038 8.12668 6.15933 10.4222 17.7634
internal_count=1000 312 688 581 525 436 322 98 114 224 89 62 98 93 76 79 55 126 219 53 64 52 56 40 107 82 44 40 83 166
is_linear=0
shrinkage=0.1
Tree=17
num_leaves=31
num_cat=0
split_feature=8 7 3 0 8 1 10 0 2 25 9 4 6 1 6 11 25 6 9 8 6 20 4 24 30 27 17 24 25 18
split_gain=12.2712 12.8986 6.40269 5.8416 5.70412 4.66438 5.12471 4.4284 4.23407 4.00041 3.87843 9.36427 5.38812 6.3215 3.4361 4.00872 3.61684 2.94696 2.67163 3.11481 4.40884 2.18337 1.97529 1.76392 1.86345 1.55984 1.53687 1.48045 0.803706 0.635859
threshold=0.36960989236831671 -0.17649734020233152 -2.0779349803924556 -1.0424582958221433 -1.6528002023696897 -0.50649863481521595 -1.0091036558151243 -1.3959577679634092 0.73140716552734386 -1.3538856506347654 0.39663147926330572 -1.803136885166168 -1.7323389053344724 -0.46048963069915766 -2.2170965671539302 2.8651051521301274 1.0000000180025095e-35 1.9626445770263674 2.3375082015991215 -0.52487343549728382 0.79199358820915233 0.22050018608570102 1.5504462718963625 0.15805819630622867 0.1522181108593941 -0.44063740968704218 -0.76643037796020497 0.16956988722085956 -0.49975097179412836 0.022797232493758205
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 2 21 8 22 25 -7 28 -4 -9 14 -12 -13 -14 -11 23 -17 -8 19 -3 29 -1 -5 26 -25 -6 -16 -10 -2 -21
right_child=7 18 3 4 5 6 17 9 27 10 11 12 13 -15 15 16 -18 -19 -20 20 -22 -23 -24 24 -26 -27 -28 -29 -30 -31
leaf_value=-0.12424523739444296 -0.05071437760526009 0.11682994173552735 0.023439507527576731 0.12230606954757646 0.023530813952696988 -0.10941075975461274 0.062505791988309481 0.07367447322177377 -0.12193986653717703 0.036789008856484347 0.15277850328341752 -0.11519555739800998 0.064210725293529933 -0.089186725710735881 -0.046536867486711986 -0.069890681743152022 0.069043514798335773 -0.051499454046482644 -0.014289287704787859 0.12893505944483116 -0.046061089034987471 -0.01754684202548935 0.027677009863476393 -0.087263217645669189 -0.00064725914635227228 0.12119935174848265 -0.12823689637098246 -0.029339991933561117 -0.11763925730578791 0.059603417419587315
leaf_weight=5.106159642338751 2.5291373580694199 16.481011204421527 5.5134466886520368 7.5906959846615774 3.1090454459190369 4.9371855407953271 5.6046211645007151 4.0863003730773917 5.1880835816264153 4.0663307607173875 3.5693646147847202 4.7918176501989409 7.6768681406974828 4.1326992884278226 3.0700504556298247 3.6141051426529875 3.8911108449101448 3.807877890765667 2.6539673656225196 3.0931114628911001 3.3999430388212204 3.0714612901210785 3.1095104813575745 6.0376029759645444 4.2198329567909241 3.4494113326072702 9.2090032100677579 2.5876658707857132 6.1768689677119237 2.3112460672855377
leaf_count=39 22 151 32 66 19 26 34 23 32 22 18 34 39 25 20 24 21 19 16 25 20 19 17 38 25 21 63 17 55 18
internal_value=0 0.0261631 -0.00278705 0.012036 0.0354245 0.00503446 -0.026897 -0.0316665 -0.0435936 -0.0217424 -0.0289258 0.00583426 -0.0257593 0.0105301 -0.0494821 -0.0611594 0.00214034 0.0163843 0.0811587 0.091177 0.0431566 -0.08417 0.0948066 -0.0822398 -0.0516301 0.0748994 -0.10781 -0.0911239 -0.0981973 0.0992845
internal_weight=0 81.0144 53.0752 44.8975 31.6083 20.9081 14.3497 67.0711 13.2892 58.3651 54.2788 20.1707 16.6014 11.8096 34.108 30.0417 7.50522 9.4125 27.9393 25.2853 8.8043 8.17762 10.7002 22.5365 10.2574 6.55846 12.2791 7.77575 8.70601 5.40436
internal_count=1000 571 341 283 202 119 79 429 81 352 329 116 98 64 213 191 45 53 230 214 63 58 83 146 63 40 83 49 77 43
is_linear=0
shrinkage=0.1
Tree=18
num_leaves=31
num_cat=0
split_feature=8 6 4 9 9 8 0 3 5 7 30 6 0 11 7 3 11 0 1 25 4 7 18 4 4 23 15 19 12 4
split_gain=10.4596 6.28903 7.32945 13.0834 9.3728 7.0822 5.70239 4.90738 5.79882 5.23404 5.86713 5.00412 5.89579 4.41822 4.35855 4.30596 4.17374 3.83232 3.06132 2.85521 3.39416 4.57969 2.38691 2.27198 1.87137 1.45707 0.556429 0.525086 0.376164 0.208452
threshold=-1.8904331922531126 2.2315721511840825 -1.0175390839576719 0.14427632093429568 -1.7243103981018064 -0.49343696236610407 -1.5799034237861631 -1.7815890312194822 -0.87118589878082264 -2.4278253316879268 0.38807439804077154 -0.60408353805541981 -1.1948255300521848 1.9994604587554934 0.81064110994339 1.4348896741867068 2.1326044797897343 0.46007883548736578 0.3707870244979859 -0.73441889882087696 1.6306554079055788 0.25074130296707159 -0.57033929228782643 -1.803136885166168 1.5504462718963625 -0.33417467772960657 0.40190966427326208 1.3435922861099245 0.22171582281589511 0.50270438194274913
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=6 2 3 5 13 28 -1 25 11 -10 19 29 -13 -4 16 18 22 -3 -14 -11 21 -21 -7 -5 27 -6 -24 -8 -2 -9
right_child=1 17 4 23 7 14 24 8 9 10 -12 12 15 -15 -16 -17 -18 -19 -20 20 -22 -23 26 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.0431961927157661 0.1238319058110581 -0.13249616896242641 -0.0025463317518052194 0.13572140118523057 -0.050853923938572278 -0.022928175540939354 0.11536345093127512 -0.092426041852938012 -0.071836467039428628 0.14329092008229441 -0.055567214569884921 -0.13703730945645487 0.12298463133361159 0.13043052637349067 0.040778720578686187 -0.068037323002512071 0.046733721857326486 -0.017420591145878694 -0.0097241233450583199 -0.070712388532502474 0.12644600902469885 0.092845575113892964 -0.13345890545203251 0.047676153495524815 0.029588126857126629 -0.13289678924638973 -0.079281774665217505 0.057589035066861582 0.068467628155879354 -0.12439162359267973
leaf_weight=3.9917711988091495 2.48240599036217 6.4759467169642431 4.8442846834659559 9.6084363758563978 3.3221813067793846 3.9141061305999765 11.481501929461958 2.8598547652363742 6.2625114396214512 3.6854804083704975 5.3332816511392585 3.4809036552905992 3.6728804931044579 5.1600456535816193 5.3634048402309409 4.1952026039361954 2.8048096001148215 5.232104517519474 3.300007700920105 3.8836981505155617 3.156420573592186 3.0614711791276932 6.1219318434596062 4.2172109782695761 4.0196817442774764 6.213101498782633 2.7461033165454865 1.8228668347001065 2.4270479828119278 7.116599015891552
leaf_count=23 23 45 32 65 23 24 118 25 37 21 36 25 19 32 36 25 17 33 21 28 18 21 41 29 23 44 18 20 19 59
internal_value=0 -0.0114181 -0.00395245 0.0303389 -0.0235198 -0.0116443 0.0645543 -0.0385674 -0.0260316 0.00750896 0.0334971 -0.0606038 -0.0234027 0.0660406 -0.0369776 0.0120153 -0.0637332 -0.0810711 0.0601785 0.06795 0.0404625 0.00138492 -0.0879732 0.108865 0.089382 -0.104312 -0.116682 0.107448 0.0964619 -0.115228
internal_weight=0 120.941 109.233 39.6855 69.5479 25.8598 21.3158 59.5436 50.0083 25.3829 19.1204 24.6254 14.649 10.0043 20.9504 11.1681 15.587 11.7081 6.97289 13.7871 10.1016 6.94517 12.7821 13.8256 17.3241 9.53528 8.86804 13.3044 4.90945 9.97645
internal_count=1000 816 738 272 466 178 184 402 335 161 124 174 90 64 136 65 100 78 40 88 67 49 83 94 161 67 59 138 42 84
is_linear=0
shrinkage=0.1
Tree=19
num_leaves=31
num_cat=0
split_feature=8 7 3 4 0 3 0 6 0 8 8 7 6 1 15 13 1 11 17 25 9 8 6 8 6 30 7 29 25 20
split_gain=9.53551 10.296 5.41519 5.30485 5.04297 4.34378 3.6182 3.53172 3.66217 3.7688 3.49489 3.31183 4.62995 5.41739 4.68222 3.97356 2.96451 3.05638 2.81643 2.49775 2.42447 2.79233 3.44687 2.04123 1.98531 1.72616 1.18934 1.04665 0.79803 0.587279
threshold=0.36960989236831671 -0.20904585719108579 1.0190169811248782 1.5504462718963625 -0.90527260303497303 -2.4990454912185665 -1.3959577679634092 0.77008655667305004 1.5622318387031557 1.701290547847748 -1.5996349453926084 0.72382399439811718 -1.868333578109741 -0.93864721059799183 -0.60756465792655934 -1.0000000180025095e-35 2.0415892601013188 1.2296399474143984 0.65595531463623058 1.0000000180025095e-35 2.3375082015991215 -0.52487343549728382 0.79199358820915233 1.4360643625259402 0.68144974112510692 -0.21610008925199506 -1.2958345413208006 -0.19384890794754026 -0.52195957303047169 0.65587282180786144
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 2 4 18 25 -6 28 11 27 -10 -7 12 26 19 -15 -13 17 24 29 -14 21 -3 -23 -16 -12 -1 -8 -9 -2 -4
right_child=6 20 3 -5 5 10 7 8 9 -11 16 15 13 14 23 -17 -18 -19 -20 -21 -22 22 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.041089906691224323 -0.043364172676169592 0.11356794165489387 0.13168426349240847 -0.027637735092789641 -0.10290375523050488 0.091791075262969299 -0.12549356939436546 -0.054793625094332755 0.062889426341094165 -0.084500193578221541 -0.075780110430592448 -0.013939242527467952 0.10829623196824552 0.049882333429194743 -0.030858345092719359 0.12225715349234162 -0.08032044082946746 0.086266084579998903 -0.00919070791354922 0.0028675949879069284 -0.011758905729667013 0.093872667929265546 -0.03858500992472496 -0.12188215233575175 0.037549983939814714 -0.1272678575336621 -0.041586614547622018 -0.1266962397993085 -0.11443791979622059 0.069802357367496917
leaf_weight=4.2424374967813492 2.2397758439183235 14.315382249653343 5.9742231182754058 6.4710994139313689 3.5455678254365912 4.7690057978034046 4.0784083977341652 3.0927241519093505 3.7536706030368796 3.2257792502641678 3.3621240928769129 4.4792710915207863 5.1467878445982933 4.3611354753375027 5.134315714240083 4.105574004352091 3.9863972067832938 4.2590803205966949 2.3241178765892974 3.9886278510093689 2.7973926812410346 4.8531296849250793 3.3007587268948555 4.7363804802298546 2.8611801490187645 5.1406614482402784 2.883833527564998 5.8610803186893454 5.3613531999289981 2.0632738582789898
leaf_count=27 21 152 47 37 26 36 29 25 19 23 22 28 29 24 32 23 23 23 18 20 18 43 20 37 19 40 19 44 56 20
internal_value=0 0.0241812 -0.0025563 0.0433992 -0.0266048 -0.00119473 -0.0288368 -0.0198761 -0.0595328 -0.00523147 0.0175505 -0.00363882 -0.0191596 0.00216724 -0.0364094 0.0511947 -0.00691972 0.0209942 0.0877636 0.0622649 0.0760326 0.0869625 0.0402527 -0.0745354 -0.0236763 -0.0883037 -0.0907384 -0.10186 -0.0934951 0.115799
internal_weight=0 74.2658 48.9992 16.8327 32.1665 22.7834 62.4487 54.8476 15.9333 6.97945 19.2378 38.9143 30.3295 23.3672 14.2318 8.58485 14.4688 10.4824 10.3616 9.13542 25.2667 22.4693 8.15389 9.8707 6.2233 9.3831 6.96224 8.9538 7.60113 8.0375
internal_count=1000 571 338 122 216 149 429 352 111 42 123 241 190 142 93 51 87 64 85 49 233 215 63 69 41 67 48 69 77 67
is_linear=0
shrinkage=0.1
Tree=20
num_leaves=31
num_cat=0
split_feature=8 4 3 0 9 7 7 9 9 18 5 30 19 15 7 9 15 6 11 26 19 14 31 13 8 31 0 6 5 29
split_gain=8.22794 8.55322 8.72736 6.67721 5.21732 4.67741 4.37093 3.48331 4.21298 3.56451 3.0786 4.21122 4.04276 3.26109 3.33921 4.51369 2.99404 2.70173 2.68262 2.60129 2.51579 2.11536 2.00794 1.93228 1.77609 1.2 1.0862 0.738621 0.571761 0.562582
threshold=0.84256586432456981 -1.6327362060546873 -1.0769400000572202 -2.2653955221176143 -0.76543956995010365 0.19911569356918338 -0.20904585719108579 1.2183623313903811 -1.6512793302536009 -0.24942214041948316 -2.1957887411117549 0.55339282751083385 0.83948722481727611 0.14212975651025775 0.81064110994339 0.39663147926330572 0.12817254662513736 -0.68222475051879872 0.40085124969482427 -0.32592111825942988 0.16267675906419757 -0.42796327173709864 0.49908749759197241 1.0069346427917483 0.36960989236831671 -0.85310545563697804 0.68990981578826915 1.7207360267639162 2.0177078247070317 -0.50869739055633534
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 4 18 -4 5 21 7 8 -5 -10 23 12 13 22 15 -15 -11 -9 -3 -20 -19 -1 25 29 27 -12 -13 -8 -6 -2
right_child=10 2 3 6 28 -7 24 17 9 16 11 26 -14 14 -16 -17 -18 20 19 -21 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.090693707116290295 -0.065042365454230103 -0.10550674255497089 -0.086457985229682122 0.073682223403565159 0.11907118994025777 0.11436143377829083 0.11710250376753953 -0.029306791978338756 -0.13284893894453903 -0.064115360994743326 -0.02933204676273234 -0.11910035124339878 0.07976423904263738 -0.0850553847290689 0.11352399979614047 0.077010018621079257 0.063231316068159793 0.03182091077219653 0.035450092900213843 -0.084158634480838032 0.13671783220898259 0.020321481362095828 0.0053883553987115196 0.0019002309211625754 0.013005863196119441 -0.10837104188771461 -0.048917922890603205 0.050175222904306843 0.057261248070375553 -0.11945732473905878
leaf_weight=3.1709974035620689 2.8266951739788047 8.1415874771773797 5.1175050102174273 3.598842471837993 11.612285133451222 3.2827657461166382 7.8479201719164866 3.1052284836769095 3.5891752317547745 4.513525389134891 2.974416643381117 3.8292928263545019 5.1644777804613105 3.8862719833850843 2.9512937963008881 3.0808605775237083 3.1240934953093529 5.1397437602281553 3.787240095436573 3.4974523894488803 4.1184479482471943 3.7417332567274588 4.0447489619255057 2.2800107300281516 2.8109148517251006 5.4231346845626902 5.199571780860424 2.0876247137784958 1.7179826498031605 5.7953724488615981
leaf_count=20 26 67 37 22 111 37 83 19 22 28 21 35 31 29 15 21 19 33 26 26 27 24 25 16 18 40 34 17 19 52
internal_value=0 0.0187589 -0.00114235 0.021086 0.0699205 0.0160728 0.0348671 0.0122152 -0.0204708 -0.0506524 -0.03333 -0.0194203 1.92605e-05 -0.0183988 0.024374 -0.0133902 -0.0120254 0.0514107 -0.066061 -0.0219752 0.0784836 -0.0306032 -0.0524953 -0.0799684 0.0831852 -0.0803754 -0.0786835 0.10304 0.111105 -0.101618
internal_weight=0 84.0051 60.4793 45.053 23.5258 10.1955 39.9355 27.1891 14.8256 11.2268 47.4561 36.5541 27.5252 22.3607 9.91843 6.96713 7.63762 12.3634 15.4263 7.28469 9.25819 6.91273 12.4423 10.9021 12.7465 8.39755 9.02886 9.93554 13.3303 8.62207
internal_count=1000 655 444 325 211 81 288 170 91 69 345 251 182 151 65 50 47 79 119 52 60 44 86 94 118 61 69 100 130 78
is_linear=0
shrinkage=0.1
Tree=21
num_leaves=31
num_cat=0
split_feature=8 4 9 3 6 9 8 3 4 31 6 11 5 8 0 0 7 6 1 5 0 3 14 11 9 23 12 5 31 5
split_gain=6.97935 7.43605 7.52251 8.2309 7.05121 4.73973 4.35368 3.02338 3.24672 2.97555 2.95816 2.77555 4.22251 2.74429 2.44577 3.19026 2.8577 4.58237 4.92304 5.2899 2.37963 1.82171 1.51378 1.29608 1.03866 0.753773 0.674422 0.573987 0.303617 0.205506
threshold=0.84256586432456981 -1.6327362060546873 -2.0369137525558467 -1.1654398441314695 1.7207360267639162 -0.76543956995010365 -0.82658091187477101 2.2420258522033696 -0.1948781609535217 0.80102509260177623 0.77008655667305004 2.4863992929458623 0.76440930366516124 -1.0591717958450315 3.099624872207642 -1.3959577679634092 0.72382399439811718 -1.5970541238784788 -0.77962037920951832 -0.87853625416755665 1.5622318387031557 1.4809150695800783 0.48551496863365179 1.6042537689208987 1.1047214865684511 0.096224114298820509 -0.23700144886970517 2.1351814270019536 0.62522068619728099 -0.27108398079872126
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 5 22 13 7 6 29 8 24 11 14 12 25 -4 15 28 17 26 -19 -20 -12 -6 -3 -9 -5 -10 -17 -7 -2 -1
right_child=10 2 3 4 21 27 -8 23 9 -11 20 -13 -14 -15 -16 16 -18 18 19 -21 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.12173952947425355 -0.11027179971226185 0.12256428370904836 -0.025086382819962956 0.042126767594388326 -0.030836247299300807 0.11633846673702153 -0.035263545690390592 0.12241253292933213 -0.1407406696769081 0.064963165958629221 -0.10918240541569586 0.02965804253660885 0.033621929532664803 -0.11968945410816401 -0.08761530622907468 -0.12338088198747894 0.070130110729473769 0.08119375900445655 0.049592260044348815 -0.11826017166595862 -0.021042647311676526 -0.13724351402478624 0.026746093971114299 0.040701856645250417 0.13006056447716116 -0.06980637346907155 -0.052020746657266852 0.051897230015873978 -0.053314501894947144 0.072468674025793656
leaf_weight=2.2181489281356339 2.466071598231792 5.1521951593458644 4.9594510197639456 3.4140625745058086 3.0733119659125796 10.797382429242132 6.2533320039510745 3.9617789536714545 3.6396225318312698 4.0638680383563033 7.3448789827525598 4.756003461778163 2.9638125263154507 8.0329971797764284 4.1267500892281523 2.8029360473155975 5.6866158843040457 6.7401622235775029 3.660100400447849 3.8551714122295344 5.2544911205768585 3.3766795545816422 2.4247674196958542 3.8061905913054943 2.2146068625152111 2.5459674522280693 2.5107452869415328 1.5851288475096215 1.508316747844219 1.3689956627786157
leaf_count=25 31 39 36 32 23 112 39 37 25 24 69 29 20 71 30 21 33 37 19 32 35 19 18 29 20 22 20 18 18 17
internal_value=0 0.017689 -0.00104931 -0.0149108 0.00868119 0.0669193 0.0151142 0.0282626 0.0104498 -0.01031 -0.0311426 -0.0323086 -0.0645199 -0.0835778 -0.01555 -0.00537567 0.00772981 -0.0104032 0.0191407 -0.0365124 -0.0724243 -0.0865422 0.0919007 0.0823755 0.0767244 -0.111544 -0.0896628 0.108089 -0.088656 0.102936
internal_weight=0 80.6083 58.3853 50.8084 37.8159 22.223 9.84048 31.3659 23.5979 17.9693 45.9562 13.9054 9.1494 12.9924 33.3569 29.2301 25.2557 19.5691 14.2554 7.51527 12.5994 6.44999 7.57696 7.76797 5.62867 6.18559 5.31368 12.3825 3.97439 3.58714
internal_count=1000 655 444 387 280 211 81 238 172 120 345 96 67 107 241 211 162 129 88 51 104 42 57 66 52 47 41 130 49 42
is_linear=0
shrinkage=0.1
Tree=22
num_leaves=31
num_cat=0
split_feature=8 4 5 7 4 8 6 6 1 11 11 9 17 2 29 3 9 3 0 12 13 30 8 10 2 21 28 8 5 11
split_gain=6.17579 4.76351 4.41919 7.19456 4.95752 5.08659 5.91485 4.49229 4.83308 5.47511 4.10357 6.12419 4.03917 3.88305 3.67237 3.99823 2.62534 2.48957 2.31024 2.13035 1.6146 2.83459 1.64794 1.60128 1.56632 1.55572 1.55026 1.19801 1.01272 1.00637
threshold=-1.8904331922531126 3.2311422824859624 -0.61190077662467945 -1.6289549469947813 -0.44345408678054804 1.4684590697288515 2.1820205450057988 -1.829513370990753 -1.4118550419807432 1.0588229894638064 2.9298924207687382 0.71947535872459423 0.045474953949451453 3.6737819910049443 -0.1444931477308273 -0.60745814442634571 -0.9361974000930785 1.6056401729583742 -1.4881429672241209 -1.0012963414192197 -1.0000000180025095e-35 0.87502706050872814 -0.2367703318595886 -1.6019649505615232 -0.026035249233245846 -1.0000000180025095e-35 0.79149156808853161 0.016810417175292972 0.37042063474655157 0.95515820384025585
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 18 4 10 5 6 14 24 23 17 11 -4 -13 19 -2 -16 -7 27 -1 -6 21 22 -21 -9 -5 -24 -20 -10 -11 -17
right_child=2 -3 3 7 13 16 -8 8 9 28 -12 12 -14 -15 15 29 -18 -19 26 20 -22 -23 25 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.0031945846544346509 0.12476706908961499 -0.099132395579474467 -0.11908341490746517 0.041001153454294226 -0.011920256257839712 -0.11571503705172653 -0.081814804443216274 0.10734582662407816 -0.023971701300368627 0.028750396795400968 0.055925103109569163 -0.069679927571811504 0.093826781260304554 0.035730278060288924 -0.067765799940200766 0.036438928491156618 0.0072627169505964442 0.031436632368520401 0.1131680107085751 0.032804087278658342 -0.11764055353280232 -0.14719307220334396 -0.0088480758820501573 0.016005305823624603 0.13245284706537039 -0.11617779842482337 0.032091081521997245 -0.10416299121971602 0.12017280374407342 0.12803265862349944
leaf_weight=3.1724927946925172 5.6439258866012123 1.7940341606736172 8.3020703643560427 2.8010383993387222 4.7327931337058606 4.3613565862178802 3.5570181012153617 5.9032602459192276 2.720514692366125 2.4323690868914118 3.6115126013755789 2.9793439283967009 3.0652779154479504 3.8469119518995276 3.175187990069392 2.7479220815002909 2.8837230652570716 2.6808703020215026 9.2251101210713369 2.7204420566558882 9.0978818275034357 2.8098117001354703 2.6949510425329253 2.8439162746071833 5.6515034362673742 2.707034714519974 3.1683318130671987 5.9101469554007053 2.4143674001097679 2.1289382353425026
leaf_count=20 40 11 65 23 35 41 26 36 25 19 18 23 20 23 20 24 23 26 117 20 82 18 22 24 46 27 36 50 23 17
internal_value=0 0.055166 -0.00924593 0.0116819 -0.0294677 0.00354984 0.033078 0.0391549 0.0177762 -0.0146355 -0.0393503 -0.0633341 0.0132357 -0.05774 0.0629172 0.0195647 -0.0667668 -0.0527388 0.0729495 -0.0722605 -0.086518 -0.0606176 -0.0306683 0.0776488 0.102147 -0.062633 0.092441 -0.0788855 0.0742918 0.0764231
internal_weight=0 17.36 104.424 51.3162 53.1079 24.4981 17.253 33.358 24.9054 16.1583 17.9582 14.3467 6.04462 28.6098 13.696 8.05205 7.24508 11.3115 15.5659 24.7629 20.0301 10.9322 8.12243 8.74718 8.45254 5.40199 12.3934 8.63066 4.84674 4.87686
internal_count=1000 184 816 398 418 191 127 272 203 143 126 108 43 227 101 61 64 101 173 204 169 87 69 60 69 49 153 75 42 41
is_linear=0
shrinkage=0.1
Tree=23
num_leaves=31
num_cat=0
split_feature=0 8 7 6 10 22 8 7 11 9 4 9 11 9 0 6 4 28 6 3 10 26 13 2 18 30 16 8 5 3
split_gain=5.47494 6.35969 7.12263 4.44669 4.23534 3.58682 4.31592 5.10999 4.90772 4.25466 9.14314 3.60615 3.15726 3.15332 2.88315 3.63586 3.54171 2.87217 2.59388 2.312 2.0907 1.83846 1.76536 1.72153 1.58631 1.46474 1.18021 1.31847 1.14641 0.543845
threshold=-3.1315989494323726 -0.82658091187477101 -1.6642212867736814 1.7042843699455263 -0.66635462641715992 1.4963601827621462 2.4709751605987553 0.72382399439811718 0.24271030724048617 -1.1903105974197385 -2.0268275737762447 -1.8577739000320432 6.7042803764343271 0.2187360227108002 0.64647096395492565 1.9243513345718386 1.8425122499465945 -0.87715819478034962 -0.60408353805541981 1.1309773921966555 -0.93400558829307545 -0.19718474894762036 0.26514133810997015 2.6751513481140141 -0.2674479186534881 -0.074140802025794969 -0.077172294259071336 0.55232834815979015 0.64191484451293956 0.70419779419898998
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=-1 2 4 5 -2 6 7 8 9 -3 -11 -10 26 22 15 16 -6 -15 -9 28 -19 -8 -13 -4 -23 -22 27 -5 -12 -28
right_child=1 3 23 12 14 -7 21 18 11 10 19 13 -14 17 -16 -17 -18 20 -20 -21 25 24 -24 -25 -26 -27 29 -29 -30 -31
leaf_value=-0.1044610555519471 -0.099156268737330355 -0.1222549860308124 0.10599963233601045 -0.0025529192729288491 0.10537126860961282 0.077700532850685733 -0.1177524725587334 0.12485029473444192 0.1086759463024507 0.12760490775456373 -0.1144029911978951 -0.11175984108643955 0.038720202204482995 -0.057474996861984318 0.11728984785986325 -0.098173661067715606 -0.060036086239216246 -0.0050291715924694743 0.018514328937125242 0.015865626217385884 0.13338708505890937 -0.094252019134004508 -0.0024289323316622254 0.028236075843621807 0.0098890820270450663 0.040292267196385997 -0.091430935859920889 -0.1053471215407418 -0.039957754164133365 -0.1547792739228401
leaf_weight=4.8183388635516158 3.7618068456649789 5.3204426392912891 12.623820699751379 2.5526367872953406 3.4041391648352128 5.4519100673496714 4.1335166953504086 4.6597904078662395 3.9658567011356309 3.3450747728347787 5.3834149204194492 2.9420690163970091 2.4231374412775031 2.8874416276812509 2.5374101325869569 2.762420617043972 2.0888327956199646 3.7234983146190599 4.5183002613484842 3.0061205700039872 3.4872749522328332 2.4237954840064049 2.9655802510678768 3.6757570207118979 3.6885239481925947 3.2794814817607509 3.0016022585332385 2.4409394860267639 3.3593896515667439 2.4707169234752655
leaf_count=48 25 45 166 21 31 31 37 42 24 23 44 22 16 22 26 18 15 27 38 25 24 20 21 38 31 26 32 20 27 15
internal_value=0 0.00439808 0.043061 -0.0102514 -0.00778435 -0.00011789 -0.0068426 0.00467428 -0.00958243 -0.0453611 -0.018257 0.0218311 -0.0641393 0.00397219 0.0240631 -0.0045914 0.0424712 0.0308434 0.072502 -0.0597858 0.055153 -0.0662419 -0.0568768 0.088463 -0.0314073 0.0882691 -0.087954 -0.0528004 -0.0857977 -0.120032
internal_weight=0 112.285 30.8542 81.4305 14.5546 68.5415 63.0896 52.8437 43.6656 20.4144 15.094 23.2512 12.889 19.2853 10.7928 8.25539 5.49297 13.3777 9.17809 11.7489 10.4903 10.2458 5.90765 16.2996 6.11232 6.76676 10.4659 4.99358 8.7428 5.47232
internal_count=1000 952 319 633 115 529 498 410 330 164 119 166 104 142 90 64 46 99 80 96 77 88 43 204 51 50 88 41 71 47
is_linear=0
shrinkage=0.1
Tree=24
num_leaves=31
num_cat=0
split_feature=8 7 3 0 6 21 1 5 9 6 6 0 0 12 30 10 1 6 8 6 13 3 22 13 9 4 9 15 8 31
split_gain=4.79142 6.61571 4.11788 4.11569 4.7947 3.97029 3.49319 3.10768 2.76577 2.7451 2.72612 2.96779 2.56139 2.43709 2.8385 2.25992 2.56338 2.83735 2.07003 2.01924 1.84843 1.71273 1.66855 1.65656 1.27828 2.40774 1.21625 0.815199 0.734757 0.331485
threshold=0.62742966413497936 -0.20904585719108579 -2.7369303703308101 0.21584135293960574 2.090929508209229 0.47412592172622686 1.2552348375320437 -1.0008283853530882 -1.6292568445205686 0.51976311206817638 0.77008655667305004 1.5622318387031557 -1.3959577679634092 0.84090170264244091 0.55339282751083385 0.82871100306510936 -0.6302025020122527 -1.7323389053344724 -0.52487343549728382 -0.85458046197891224 0.44129484891891485 1.0971081256866457 0.26117700338363653 1.3032695651054385 0.31265056133270269 -0.14276775717735288 0.34155946969985967 0.34851495921611791 1.9199492931365969 0.62522068619728099
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 2 -1 4 5 9 24 18 -8 21 12 -12 29 14 15 16 19 -18 26 -14 -16 -4 -7 -9 25 -5 -3 -26 -19 -2
right_child=10 7 3 6 -6 22 8 23 -10 -11 11 -13 13 -15 20 -17 17 28 -20 -21 -22 -23 -24 -25 27 -27 -28 -29 -30 -31
leaf_value=-0.11317312758767809 -0.10974175191366442 0.11430163087253345 -0.10808513532304112 0.097637285085163103 -0.12673099248559858 0.012755861671304519 0.059816171402328144 0.10388569999741559 -0.063786387461327185 0.037889851076871921 -0.10796341002451373 -0.010343170058382997 -0.00014526860625723606 -0.063748600373387435 -0.094863063570683859 0.10572003089413984 0.046167890061014946 -0.03661126852847521 -0.044081517636049956 0.103482849016977 0.016559461366202639 -0.017306391501049345 0.11624481735564883 0.001040569793706359 0.12635431811811051 -0.043299441466538434 0.00021817360182675968 0.056300765583247182 -0.11632274377065582 -0.047721495406016536
leaf_weight=3.2430082522332659 2.4252344630658622 2.043133314698935 4.7913936227560097 2.6047233417630196 3.8765092566609374 3.4056123085319987 2.9703056588768959 12.195765018463133 4.6357301957905284 3.3810634426772586 6.9791832305490953 5.6236272379755974 3.4045920595526695 5.0230077020823947 3.4603617861866942 3.382529951632022 3.3427443504333532 2.419759579002859 3.5804031938314465 4.1998831033706701 2.613266184926033 3.6705225370824337 2.8715905025601387 1.7969340868294228 4.2274334356188774 2.267307288944723 1.7221922092139719 2.7363503165543079 2.2148358784615993 1.3367993757128713
leaf_count=30 34 27 43 27 30 29 19 163 38 30 77 39 23 38 24 22 23 19 26 26 21 27 20 20 35 14 15 22 21 18
internal_value=0 0.0172207 -0.00465531 0.00383737 -0.0257911 -0.0041967 0.0373593 0.0630278 -0.0155172 -0.0382753 -0.0247056 -0.0644033 -0.00991384 -0.000178713 0.0125744 0.031629 0.0155452 -0.0240555 0.010357 0.0570876 -0.0469219 -0.0687081 0.0600983 0.0906784 0.0713392 0.0320493 0.0621219 0.0988275 -0.0747047 -0.0877035
internal_weight=0 66.02 44.6816 41.4385 21.9967 18.1202 19.4419 21.3384 7.60604 11.843 46.4258 12.6028 33.823 30.061 25.038 18.9643 15.5818 7.97734 7.34573 7.60448 6.07363 8.46192 6.2772 13.9927 11.8358 4.87203 3.76533 6.96378 4.6346 3.76203
internal_count=1000 615 364 334 179 149 155 251 57 100 385 116 269 217 179 134 112 63 68 49 45 70 49 183 98 41 42 57 40 52
is_linear=0
shrinkage=0.1
Tree=25
num_leaves=31
num_cat=0
split_feature=0 8 7 5 11 9 7 13 5 7 3 6 6 8 14 17 7 17 6 1 7 0 0 5 29 0 16 13 4 11
split_gain=4.46525 5.02088 5.81699 3.72046 5.82294 4.60656 3.93071 3.42454 3.33264 3.22812 3.0715 2.97773 3.77739 3.45623 2.92821 2.85615 3.06228 2.78887 2.68148 2.27463 2.24404 1.96119 1.88431 1.77028 1.74927 4.16026 1.11921 0.880863 0.597152 0.294951
threshold=-3.1315989494323726 -0.82658091187477101 -1.6642212867736814 -0.61190077662467945 0.10654639080166818 -0.60508364439010609 0.28290793299674993 1.3883623480796816 0.56105643510818493 -1.4436311721801756 -0.94009315967559803 -1.7323389053344724 1.7042843699455263 1.0175021886825564 0.71469354629516613 0.49918510019779211 -1.6289549469947813 0.29379040002822882 -1.1482236981391905 0.6997278928756715 -1.0066578388214109 1.3857911825180056 -0.43635436892509455 0.72707492113113414 -0.47604635357856745 -0.10111981630325316 0.44526234269142156 -0.37099200487136835 -0.29410147666931147 -2.1716114282608028
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=29 2 8 7 5 6 -5 10 19 18 27 -12 13 17 26 16 -6 28 -7 22 -9 -18 -2 -11 25 -4 -15 -3 -13 -1
right_child=1 3 24 4 15 9 -8 20 -10 23 11 12 -14 14 -16 -17 21 -19 -20 -21 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.069945267909128458 0.000193788582284352 -0.049309367127451618 -0.075245749736506939 -0.11662425551206533 -0.044829341143195374 -0.10299193250144628 0.0078749667759020151 0.098844772844438364 -0.084113787467519463 0.0078307203013697254 -0.11009772996607647 0.12607378788047671 -0.10091738474527785 -0.119619005602176 0.029985873476860127 0.13339847818915496 0.0049981016750930505 -0.031827785152516747 0.03844504207525825 -0.028497840364942136 -0.032872027241241573 0.11219642438659419 0.12161772855268826 0.11038839782307537 0.10913085949236234 0.11050864895492599 -0.022915978004573963 -0.11654569832566375 0.06044910409079654 -0.1227494729087608
leaf_weight=1.8669797815382478 2.2410825267434111 2.9360643625259399 2.0089955031871796 6.3385890945792198 3.3061751574277869 2.7905479297041893 4.2270999103784561 3.0785720944404602 3.9671820476651183 2.8603768199682236 3.5145002380013457 3.038640633225441 3.8015473522245875 3.1819443404674512 4.1357445083558559 4.0849923342466345 2.7577932998538008 2.4500901773571968 2.5795252434909344 4.3491764962673187 2.2306371703743926 4.4775673188269138 2.9740837365388897 4.0892735049128532 9.5473335385322553 3.0154299587011328 1.9183862321078775 5.7929674498736841 2.5503976047039121 2.4407347217202178
leaf_count=24 18 21 18 65 22 25 36 19 29 20 42 25 28 35 30 27 25 17 19 38 21 30 30 37 138 48 15 56 18 24
internal_value=0 0.00403045 0.0401544 -0.00930265 0.0131291 -0.0183685 -0.0668149 -0.0310854 -0.00705769 0.0231802 -0.0429705 -0.0248815 -0.0106719 0.0091873 -0.0325429 0.0624119 0.0349036 0.0571307 -0.0350524 0.0249043 0.0435046 0.0713373 0.0694389 0.0681771 0.0839961 0.0362355 -0.0832461 -0.0939303 0.0961278 -0.0998639
internal_weight=0 104.245 28.1033 76.1414 37.5119 22.8854 10.5657 38.6295 13.5315 12.3197 33.3203 24.5913 21.0768 17.2752 9.23608 14.6265 10.5415 8.03913 5.37007 9.56434 5.30921 7.23536 5.21517 6.94965 14.5718 5.02443 5.10033 8.72903 5.58904 4.30771
internal_count=1000 952 319 633 306 202 101 327 115 101 287 210 168 140 80 104 77 60 44 86 40 55 48 57 204 66 50 77 43 48
is_linear=0
shrinkage=0.1
Tree=26
num_leaves=31
num_cat=0
split_feature=8 5 11 4 9 7 5 1 4 11 6 8 20 1 7 5 1 4 11 0 27 11 1 2 28 18 4 20 2 4
split_gain=4.02404 3.51226 4.17689 4.57091 7.35477 4.11367 3.51427 4.52521 3.38438 2.93289 3.16136 2.89243 2.56594 2.56959 2.50365 2.49345 2.17785 2.03496 2.46355 2.01193 1.99136 1.98196 1.76948 1.76289 1.65413 1.441 1.43232 1.30144 1.16422 1.12668
threshold=-1.8904331922531126 -3.4838403463363643 4.6643924713134775 -1.0175390839576719 0.14427632093429568 0.88489025831222545 -0.89550951123237599 -3.659884929656982 3.2311422824859624 0.10654639080166818 0.32099404931068426 2.8510276079177861 0.60165283083915722 1.1597656607627871 -1.2958345413208006 1.4445725679397585 -0.20998486876487729 0.90728291869163524 0.91672056913375866 -1.5799034237861631 0.22275302559137347 -0.43148031830787653 1.0285736322402956 -0.23842298984527585 0.79149156808853161 0.26834392547607427 -2.1853830814361568 -0.52321508526802052 1.7490286231040957 -1.6327362060546873
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=8 27 3 4 5 21 17 -8 19 15 14 -7 13 -4 -11 23 -17 18 -5 -1 -20 25 -23 -9 -21 26 -3 -2 -12 -6
right_child=1 2 12 6 29 11 7 9 -10 10 28 -13 -14 -15 -16 16 -18 -19 20 24 -22 22 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.01066094889755601 -0.016942396934092968 0.014819885930214771 0.078092957521398387 -0.12015919877903564 0.11419798388609215 0.12168922410210614 0.1186377992317965 -0.027189170516987157 -0.086324088508550043 -0.015842480365081131 -0.10593891058727267 -0.039841940673543047 0.12175313376644792 -0.035546591856823566 0.096080266549434593 0.079808726046003239 -0.061880993417804134 -0.11189815584705823 -0.043299040660922432 0.11084347711034799 0.083647251509145495 0.076301610984079338 -0.043473820099446892 -0.11826556813519361 0.018554033639288679 -0.1142471799861671 -0.089953677397340059 -0.11310239902999918 -0.0064607853858103127 0.032108741019280573
leaf_weight=2.6804982572793952 2.1383649520576 2.7509362809360045 3.9587942659854871 2.477359980344767 6.834592055529356 3.1252757944166651 2.443535272032018 3.3215173147618726 1.638457916676997 3.4916625246405593 2.281836822628974 1.7178681641817091 3.7013612538576126 4.0005548670887947 4.6741846911609164 2.0943707302212715 2.2504560612142077 6.7674039006233215 3.6287210322916508 7.4582399278879148 1.8737588077783582 2.1752767749130753 2.8486416302621365 5.9010742008686234 2.6258159801363945 4.0634141080081463 2.482001680880785 4.1176573783159247 2.4286191537976265 2.2134457528591138
leaf_count=18 18 23 25 29 67 38 18 28 11 23 18 13 23 34 39 16 25 62 35 119 17 15 29 63 36 39 26 46 24 23
internal_value=0 -0.00796423 -0.00255004 -0.0115601 0.0198091 -0.0152752 -0.0318415 -0.0115644 0.048976 -0.0235958 0.0105895 0.064394 0.0529642 0.0209751 0.0482228 -0.0560397 0.00641881 -0.0715607 -0.037352 0.0663431 -7.00348e-05 -0.0422195 0.00838704 -0.0854644 0.086812 -0.0695682 -0.0348746 -0.080234 -0.0546499 0.0941163
internal_weight=0 89.7627 83.5067 71.846 28.2115 19.1634 43.6345 28.8873 14.403 26.4437 12.8763 4.84314 11.6607 7.95935 8.16585 13.5674 4.34483 14.7472 7.97984 12.7646 5.50248 14.3203 5.02392 9.22259 10.0841 9.29635 5.23294 6.25602 4.71046 9.04804
internal_count=1000 816 752 670 273 183 397 254 184 236 104 51 82 59 62 132 41 143 81 173 52 132 44 91 155 88 49 64 42 90
is_linear=0
shrinkage=0.1
Tree=27
num_leaves=31
num_cat=0
split_feature=1 0 8 8 7 13 6 8 3 26 11 3 31 17 4 9 11 3 5 7 0 9 4 4 5 20 20 20 6 3
split_gain=3.73628 4.35109 4.60068 2.81109 4.01501 3.8042 3.4962 4.30227 3.19717 2.47334 2.4301 2.71714 2.55623 2.90692 2.92994 3.40919 3.60905 2.10325 2.0613 1.79494 1.74896 1.71168 1.57655 1.46123 1.45692 1.30549 0.676169 0.483728 0.475246 0.311273
threshold=4.1692261695861825 -3.1315989494323726 -1.8904331922531126 2.4709751605987553 0.25074130296707159 1.3883623480796816 -0.69330593943595875 0.72140043973922741 -1.7318238019943235 -0.19718474894762036 4.6643924713134775 1.0971081256866457 1.3691098093986513 0.75022095441818248 -0.53900733590126026 0.39663147926330572 -0.13035058230161664 -0.98315218091011036 1.4048260450363161 -1.6642212867736814 -1.5799034237861631 1.4647815227508547 1.2886417508125307 -1.5022985935211179 -0.63844007253646839 -0.13627474755048749 0.25876171886920935 -0.45171776413917536 0.37987798452377325 -2.4266439676284786
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 -1 20 4 5 8 -6 17 23 -5 12 -12 13 14 15 16 -10 -8 21 -7 -3 26 -22 -4 -11 -23 -16 -19 -15 -25
right_child=-2 2 3 9 6 19 7 -9 10 24 11 -13 -14 28 18 -17 -18 27 -20 -21 22 25 -24 29 -26 -27 -28 -29 -30 -31
leaf_value=-0.10406973376518387 -0.12198287598153477 -0.018239124260575636 -0.0075480134959803706 -0.11437698605422403 0.11295646442748453 0.10551941507461118 -0.032753739193541055 -0.06489512041053859 -0.082889877398819992 -0.058140573271253486 0.12406184579438631 -0.0039109976864720244 0.050899492683028263 -0.12059336496135485 -0.12931002523511456 0.10641673836196973 0.065528200802194514 0.049671918860685248 0.038308613963781268 0.0021729654079746569 0.10920425740753739 -0.06837661840199015 0.024953905994801892 -0.12068551695788421 0.034499023045985865 0.040170115387377893 -0.059737443866376153 0.11503652414395749 -0.059517256563366649 -0.071963232374679151
leaf_weight=3.6547136586159459 2.4527888353914014 2.1183300763368598 2.253821574151516 3.7786197699606419 5.6070631928741959 3.9890340603888035 1.7068558037281043 4.1254600957036018 3.6071711704134941 3.5201342366635799 2.7952961623668662 4.0818414762616158 4.7071701586246482 3.1187243536114684 3.4109774380922477 3.5146848671138287 3.0018637031316704 1.6375658661127097 2.3868545182049266 2.9040461368858805 6.8034856095910055 2.403929129242897 3.2976299300789833 3.5716102086007595 3.2789131104946119 2.0553364008665085 2.3658728376030922 3.6685438007116282 2.1539017707109451 2.0719155594706455
leaf_count=47 21 15 17 41 72 26 12 37 39 38 17 26 39 27 27 37 26 18 18 27 133 23 24 38 23 17 23 44 22 26
internal_value=0 0.00297962 0.00714449 -0.00141331 0.00573856 -0.00744282 0.0485554 0.016136 -0.017517 -0.0495129 -0.0059318 0.0481051 -0.0172871 -0.0287423 -0.0132348 0.0268409 -0.0154775 0.0638035 -0.0453758 0.0619797 0.0643746 -0.0648893 0.0816997 -0.0756147 -0.0134641 -0.0183459 -0.100817 0.0948638 -0.0956434 -0.102798
internal_weight=0 97.5914 93.9367 81.7172 71.1395 54.3941 16.7455 11.1384 47.501 10.5777 39.6036 6.87714 32.7265 28.0193 22.7467 10.1237 6.60903 7.01297 12.623 6.89308 12.2194 10.2361 10.1011 7.89735 6.79905 4.45927 5.77685 5.30611 5.27263 5.64353
internal_count=1000 979 932 760 658 475 183 111 422 102 341 43 298 259 210 102 65 74 108 53 172 90 157 81 61 40 50 62 49 64
is_linear=0
shrinkage=0.1
Tree=28
num_leaves=31
num_cat=0
split_feature=8 7 3 0 6 4 11 1 2 0 3 2 6 11 4 0 25 9 28 29 8 3 11 15 25 13 18 15 27 26
split_gain=3.2191 4.29651 2.79925 2.60999 2.76795 2.90082 2.60127 2.31032 2.62252 2.28484 2.18742 2.16766 2.03518 2.45519 2.91127 2.44489 2.26565 2.12303 2.0301 1.90813 1.68719 1.58073 2.72472 1.31479 1.2512 1.14534 1.13266 1.04243 1.01384 0.542763
threshold=0.62742966413497936 -0.86519092321395863 -2.7369303703308101 0.16223448514938357 2.090929508209229 -0.86077696084976185 -0.95134982466697682 -0.50649863481521595 1.6277096867561343 -2.9661453962326045 2.0606999397277836 1.3687226772308352 0.77008655667305004 -5.2955746650695792 -1.1332715749740598 1.5622318387031557 -0.61387494206428517 -0.9361974000930785 -0.60353773832321156 -1.0000000180025095e-35 -0.45424538850784296 0.99575218558311474 0.15818990021944049 -0.96562194824218739 -0.013136144261807202 0.55809754133224498 0.25012546032667166 0.14212975651025775 0.18127222359180453 0.041791640222072608
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 2 -1 4 5 -4 -7 26 -9 -3 18 -8 13 -2 17 -14 27 28 20 -10 -11 22 29 -20 -23 -25 -5 -16 -15 -18
right_child=12 9 3 7 -6 6 11 8 19 10 -12 -13 15 14 16 -17 21 -19 23 -21 -22 24 -24 25 -26 -27 -28 -29 -30 -31
leaf_value=-0.10701837535237811 -0.10520888486781717 -0.085555815192350892 0.079153510063350696 0.12180220007863383 -0.1207406256835953 -0.099958157322894009 0.066625926080390238 -0.058475532686229514 0.10193513469159086 0.049252708082766938 0.1139642894880514 -0.046802818115454059 -0.10324616151876775 -0.047471010028764082 -0.011663117856237802 -0.0078274345315718849 -0.11417392490020739 0.099303844136373887 -0.012019891991587685 -0.014084308489062242 -0.079236957344865022 -0.029703533108522997 0.036859726854386418 0.097082097345235929 -0.12968986206766261 0.020124567672497155 0.027097851186362539 0.0804329702083319 0.047079965750937314 -0.037579807857163881
leaf_weight=2.5854780636727801 2.3883102945983437 1.1834458298981214 2.9535089842975166 3.2012700140476236 2.6211339160799971 3.7414247319102296 3.0422857739031324 4.106881190091368 2.9623196460306636 1.8579733632504938 5.136835452169179 3.7757588364183903 5.7690992839634401 2.229505892843008 2.387464087456471 5.0235835388302803 2.0288611948490094 4.1964582875370997 2.0778719969093826 2.7184545993804941 2.2711561620235488 2.1753812432289124 4.4125204458832741 6.9557041339576173 2.947028348222374 2.6786228902637959 2.0856464821845293 2.5329836644232273 2.3081116378307343 1.7006790749728677
leaf_count=28 24 10 30 29 25 43 23 36 27 34 84 33 77 23 22 39 23 32 27 21 20 21 34 91 33 38 16 21 17 19
internal_value=0 0.0154137 -0.0070259 0.00125796 -0.0266952 -0.00845309 -0.0329569 0.0311759 0.00240352 0.0496318 0.0572582 0.00381032 -0.0217086 -0.00803739 0.000583875 -0.0588324 -0.0222072 0.0480364 0.0388702 0.0464157 -0.0214208 -0.0437052 -0.0163239 0.0601257 -0.0872277 0.0756857 0.0844421 0.0357468 0.000623437 -0.0792468
internal_weight=0 55.9558 33.7942 31.2087 16.1341 13.513 10.5595 15.0746 9.78766 22.1616 20.9782 6.81804 40.1 29.3073 26.919 10.7927 18.1849 8.73408 15.8413 5.68077 4.12913 13.2645 8.14206 11.7122 5.12241 9.63433 5.28692 4.92045 4.53762 3.72954
internal_count=1000 615 311 283 154 129 99 129 84 304 294 56 385 269 245 116 173 72 210 48 54 130 76 156 54 129 45 43 40 42
is_linear=0
shrinkage=0.1
Tree=29
num_leaves=31
num_cat=0
split_feature=1 0 8 5 7 11 9 4 5 6 1 6 4 3 7 8 0 6 9 4 6 18 25 7 5 25 23 3 8 5
split_gain=3.24578 3.40562 3.65181 2.52328 3.96258 3.33911 3.73858 2.95386 3.48007 2.8814 2.81686 2.61435 2.59946 3.38923 3.68245 2.5992 2.57889 2.41357 3.67548 2.26768 1.92722 1.71794 1.54677 2.41183 3.55182 1.36602 0.930239 0.911469 0.625779 0.51421
threshold=4.1692261695861825 -3.1315989494323726 -1.8904331922531126 -0.61190077662467945 -2.499995112419128 0.10654639080166818 -2.2834039926528926 -1.0345802903175352 0.97322815656661998 -1.7323389053344724 -1.1931331753730772 2.7370777130126958 1.4291652441024782 -0.27606296539306635 -1.3431769609451292 1.0356987714767458 2.3296927213668828 0.32099404931068426 -1.9994176626205442 -1.6327362060546873 0.54890441894531261 0.091590221971273436 -0.60713538527488697 -0.7118559181690215 -0.74314984679222096 0.10375598073005678 -0.045334011316299432 0.24225854873657229 1.7433168888092043 -2.1659255027770992
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 -1 22 9 10 6 28 26 27 -4 -5 12 13 16 -15 25 19 -7 -19 -11 -17 -14 -3 24 -24 -16 -8 -9 -6 -18
right_child=-2 2 3 4 5 17 7 8 -10 11 -12 -13 21 14 15 20 29 18 -20 -21 -22 -23 23 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.098542276660663009 -0.11720870128380401 0.11902143303744801 -0.1038038486245446 0.028631441357649896 -0.033105941751248154 0.11630776510503314 0.039882610907624787 -0.11614255400147784 0.064380853501979349 0.07962254308928092 -0.096243778629851048 -0.11162532816822431 -0.0060488800114339644 0.12272354591453256 0.0096784292048190702 0.013641381068036279 -0.11828670882886182 0.10836505855742412 -0.061282540183380918 -0.040629625988297585 -0.11213099403600915 -0.10352441986806145 0.097141363221518062 0.10637958371941601 -0.077632820387532664 0.11054938460229069 0.11393839153234522 -0.022196326278390147 -0.11433973674805052 -0.051464744792960926
leaf_weight=3.1926630977541199 2.307515885680913 3.1914863232523203 3.635051514953374 2.975316321477294 1.6453281044960051 6.4776835255324823 3.2593105286359805 2.2707370184361935 3.2730860207229853 2.7617151364684105 4.5979216732084742 2.373673614114522 3.5384439080953589 4.1979775633662877 2.6963399033993589 2.8332436010241508 2.4382263496518135 2.5797394383698702 2.5290840715169907 3.6285937279462761 2.1374242082238197 3.6973442062735558 1.9447623491287229 3.0340424440801126 2.8917778171598911 2.6738983560353518 3.5368172638118267 1.8941920436918733 2.2384709529578686 2.1823489405214787
leaf_count=47 21 47 52 24 21 55 39 25 33 21 44 25 30 39 22 24 22 21 25 38 30 33 27 72 26 29 42 20 27 19
internal_value=0 0.00289731 0.00661414 -0.00119254 0.0173868 0.0338493 0.00703742 0.0307655 -0.0127788 -0.0190455 -0.0471837 -0.0102825 -0.00294531 0.0120395 0.0437353 0.0116694 -0.0298121 0.0757752 0.0243823 0.01134 -0.0404417 -0.0558569 0.0602993 0.0364878 -0.0073565 0.0599031 0.0784225 -0.0734162 -0.0799259 -0.086726
internal_weight=0 90.3267 87.134 76.072 37.2777 29.7044 18.1179 14.2341 7.43802 38.7943 7.57324 35.1592 32.7856 25.5498 14.5389 10.3409 11.0109 11.5865 5.10882 6.39031 4.97067 7.23579 11.0621 7.87058 4.83654 5.37024 6.79613 4.16493 3.8838 4.62058
internal_count=1000 979 932 760 376 308 207 159 78 384 68 332 307 244 144 105 100 101 46 59 54 63 172 125 53 51 81 45 48 41
is_linear=0
shrinkage=0.1
Tree=30
num_leaves=31
num_cat=0
split_feature=1 0 8 2 5 1 7 20 4 9 9 11 5 4 25 4 6 23 8 22 29 4 28 0 6 1 20 4 13 6
split_gain=2.74189 2.9796 4.51562 2.21561 2.7913 2.21573 3.83373 2.5815 2.45421 3.79094 2.95586 2.88991 2.23064 1.93899 2.04561 1.90731 1.67743 1.52476 1.39284 1.37685 1.34358 1.31809 1.31408 1.55091 1.17032 1.15772 1.11919 0.859766 0.343686 0.282317
threshold=4.1692261695861825 -1.608339190483093 -1.9779059886932371 4.188784360885621 -2.2353143692016597 -0.93864721059799183 0.84199178218841564 1.5279591083526614 -0.44345408678054804 0.12808340787887576 -1.7243103981018064 -1.0000000180025095e-35 -0.95051786303520192 -0.97413113713264454 -0.80188676714897145 0.45788866281509405 -1.5804541110992429 -1.0095380544662473 2.5378247499465947 0.88484010100364696 0.74896043539047252 1.7364857196807864 -0.62721845507621754 2.3886673450469975 -0.47526878118515009 0.20622190833091739 0.37240926921367651 -2.3924001455307002 -0.53511899709701527 -0.35840791463851923
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 13 21 4 22 12 7 8 9 11 -10 27 15 -1 -15 -6 -14 -18 28 -16 -19 -3 -4 25 26 -24 -12 -7 -8 -29
right_child=-2 2 3 -5 5 6 18 -9 10 -11 24 -13 16 14 19 -17 17 20 -20 -21 -22 -23 23 -25 -26 -27 -28 29 -30 -31
leaf_value=0.040921383670639827 -0.10977256741022126 0.10565864425587455 -0.1069024999827233 0.089615730248260181 0.035137951701449809 -0.018937029407012178 0.052160501923724012 0.086260372902185203 0.037992875554881025 0.094680522300956138 -0.088863253197690353 0.033868457567810144 -0.023159568996495808 0.018467090039126135 -0.11244285933116288 -0.0905751588181051 0.0035828048041291745 0.11500356654072322 -0.0092421288129808446 -0.0047040693293754245 0.030739766709763628 0.012744127139931906 -0.049459043590548202 -0.10223906389800914 -0.11002626745876234 0.042660220162360062 0.016248017714026057 -0.068370188988685174 0.11417587596802299 -0.12082935688314048
leaf_weight=2.1730748973786822 2.2220608107745639 5.8586697317659846 2.8880337532609701 2.7425367478281251 2.5738377533853081 1.9448624551296179 1.2766565382480619 2.0050343945622435 2.4175016954541197 3.149782743304967 2.2840532995760432 4.1543333102017677 2.2992411963641652 2.2922215014696112 5.6967747490853062 2.272383239120245 2.6501743905246267 6.3375301007181282 1.8059563860297201 1.4980751816183326 2.6977493762969971 2.0649101212620735 2.5370754711329919 2.0292730964720249 6.7150932513177528 2.9512681793421507 1.8203179240226743 1.7428128793835584 2.9786695335060349 2.4938143994659239
leaf_count=21 21 134 31 24 28 21 20 15 18 37 23 39 23 25 86 15 24 56 15 19 24 17 28 22 67 34 17 21 46 29
internal_value=0 0.00273042 0.0100698 0.00159949 -0.0021707 0.00702741 -0.00792844 -0.0231768 -0.031388 -0.00136356 -0.0619765 -0.0306324 0.0346574 -0.0442836 -0.0638003 -0.0238087 0.0549181 0.0702807 0.0643416 -0.09001 0.0898441 0.0814448 -0.0495681 -0.027542 -0.0843136 7.66031e-05 -0.0422457 -0.0739809 0.0955704 -0.0992493
internal_weight=0 86.3517 74.6916 66.768 64.0255 53.6198 34.7889 28.7276 26.7226 13.4856 13.237 10.3358 18.8309 11.6601 9.48707 4.84622 13.9847 11.6855 6.06128 7.19485 9.03528 7.92358 10.4057 7.51762 10.8195 5.48834 4.10437 6.18149 4.25533 4.23663
internal_count=1000 979 828 677 653 538 368 287 272 147 125 110 170 151 130 43 127 104 81 105 80 151 115 84 107 62 40 71 66 50
is_linear=0
shrinkage=0.1
Tree=31
num_leaves=31
num_cat=0
split_feature=8 7 13 3 24 3 0 1 2 4 1 8 5 13 2 5 19 15 25 31 19 8 3 5 7 2 17 9 29 3
split_gain=2.48837 3.16244 2.375 2.51054 2.07206 2.31965 2.40153 2.03799 1.76357 2.57083 1.70308 1.64767 1.63635 1.59479 1.58676 2.20338 1.91303 2.00033 2.2608 1.86246 1.8617 1.29991 1.26062 1.17841 1.11808 0.975232 0.78299 0.694589 0.574518 0.411726
threshold=0.62742966413497936 -0.20904585719108579 1.4291877746582033 -2.7369303703308101 1.6863377690315249 2.6009738445281987 -0.90527260303497303 -1.6409657001495359 1.6277096867561343 0.74304640293121349 1.5725554227828982 -1.0591717958450315 -1.0008283853530882 1.3032695651054385 3.0034747123718266 -2.1957887411117549 0.82704880833625805 1.1267175674438479 -0.15786706656217572 0.59073963761329662 1.228871285915375 -0.52487343549728382 0.587295562028885 -0.74314984679222096 0.88489025831222545 0.97792845964431774 -0.72679358720779408 0.34155946969985967 0.5306413769721986 -0.19423490762710569
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 2 3 -1 5 6 22 -8 9 11 29 -9 21 -14 15 26 17 18 19 23 -18 27 -5 -17 28 -7 -2 -3 -20 -10
right_child=14 12 -4 4 -6 25 7 8 10 -11 -12 -13 13 -15 -16 16 20 -19 24 -21 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.11624041252769202 -0.021111217540461122 0.1100421170223293 0.078675324463742147 -0.1118225491701105 0.098989527678206513 0.12560524711985857 0.081475669439268275 0.08363185717337332 0.08877480513684792 -0.11919152567415575 -0.035342075186998438 -0.053209140890755957 0.096407360080679558 -0.019525372117916662 0.039920531297239492 0.026266235042744581 0.086803665292640234 0.044541456573585644 -0.1125352393205878 0.064572596972812965 -0.029851529594264077 -0.038165584624567928 -0.022233144552507859 -0.073595602566381396 -0.020341061425982739 0.02037146142952953 -0.10615047158633109 0.0073793287774446704 -0.050012246574985868 0.026869976954713676
leaf_weight=2.0970273353159419 1.4615087546408174 1.3503261338919403 3.2975010555237523 3.6776924189180198 1.8661131709814061 1.5204397700726979 2.5062522478401759 1.516060525551447 2.1940832585096359 3.1689141131937504 3.5202628038823605 2.0969715304672616 7.8707610201090557 1.3972002882510421 3.9366999771445981 1.8494494594633577 3.1871929317712775 3.1361196450889102 5.1145732048898953 2.7427082546055317 2.3968593291938305 2.768201880156993 2.7413639836013317 3.2726971488446059 2.8292143158614635 2.0927681922912598 4.177576193585991 1.2872787024825809 2.0622891001403332 2.1052801515907049
leaf_count=28 19 27 31 47 16 17 27 26 19 26 32 21 163 20 36 18 24 29 56 25 28 26 27 37 24 26 66 15 23 21
internal_value=0 0.0145576 -0.00202186 -0.0105772 -0.0029382 -0.0099466 -0.0214036 -0.00183317 -0.0161325 -0.05345 0.0162328 0.00421056 0.053426 0.0789298 -0.0200124 -0.0273328 -0.0152922 -0.029121 -0.0420478 -0.00192937 0.0367314 0.00970104 -0.0735619 -0.0375386 -0.0735812 0.0646539 -0.0841104 0.0599377 -0.0945691 0.0584617
internal_weight=0 49.0745 34.4007 31.1032 29.0062 27.1401 23.5269 17.1078 14.6016 6.78195 7.81963 3.61303 14.6738 9.26796 36.1669 32.2302 26.5911 21.0071 17.8709 7.86485 5.58405 5.40581 6.41906 5.12215 10.0061 3.61321 5.63908 2.6376 7.17686 4.29936
internal_count=1000 615 364 333 305 289 246 172 145 73 72 47 251 183 385 349 264 212 183 80 52 68 74 55 103 43 85 42 79 40
is_linear=0
shrinkage=0.1
Tree=32
num_leaves=31
num_cat=0
split_feature=1 0 8 7 11 6 9 25 22 9 6 7 6 8 11 9 5 1 6 1 7 5 10 30 4 10 5 7 2 16
split_gain=2.28286 2.6769 2.90906 4.26784 2.66919 2.8013 2.39872 2.14779 2.12843 2.48285 4.33427 2.13843 2.64853 2.07473 2.00335 2.02864 1.97179 2.0639 1.9364 2.21166 1.65671 2.03625 1.44746 0.993464 0.801411 0.735336 0.697379 0.570556 0.398998 0.14786
threshold=4.1692261695861825 -3.1315989494323726 -0.94087243080139149 -1.6642212867736814 4.8875966072082528 1.7042843699455263 0.24192273616790774 -0.78593450784683216 1.4963601827621462 1.2301043272018435 -1.7323389053344724 0.72382399439811718 -2.3307499885559078 2.067369937896729 -1.0000000180025095e-35 -1.8577739000320432 0.56105643510818493 0.47342592477798467 0.12518543004989627 -0.46048963069915766 -1.3801147341728208 -0.52954918146133412 -0.69993299245834339 -0.08159421011805533 -2.2204345464706416 -3.6773331165313716 0.049624711275100715 -0.68678399920463551 1.9683957099914553 1.0472776889801028
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 -1 3 7 5 8 -6 -3 9 11 26 14 -13 -14 18 -16 17 -9 29 -20 -12 -22 25 -8 -7 -17 -11 -5 -26 -4
right_child=-2 2 4 27 6 24 23 16 -10 10 20 12 13 -15 15 22 -18 -19 19 -21 21 -23 -24 -25 28 -27 -28 -29 -30 -31
leaf_value=-0.094630089661201386 -0.10377438308199338 0.071523895495285691 -0.11335029910510644 0.054084072049447388 0.098672660177568455 -0.015441805232838585 -0.057155261849244292 0.068393848130714238 0.070040627838337352 -0.11138351235944317 0.11517961102429525 0.12141829549132295 0.034067131723872296 -0.093035119464688731 0.070017420979413186 -0.025164711658124773 -0.11890980349556382 -0.063978631931298643 0.031392307660688901 -0.09998088636409036 -0.046937016524666697 0.097412040764068397 0.010784646417568431 0.039793479605358367 -0.11480928559129396 -0.10839963488146401 -0.014039981817195174 0.10926229937965691 -0.054148920534764888 -0.077383681190842504
leaf_weight=2.7382671106606713 2.0717228315770617 2.5941162444651127 5.1229000594466907 2.7429662961512804 4.383110933005808 1.7253059856593607 2.0376320835202932 2.2677066270262021 3.543282961472868 1.7864566296339037 3.9377502202987671 1.9404014330357311 3.3478157725185147 2.0835430789738894 2.4656763561069956 1.8439458534121502 1.9967442676424978 2.4508314579725266 3.0415403693914413 2.2144499234855175 2.0084926169365649 1.9033141881227473 3.9109157286584439 2.1962391473352909 3.2717787493020287 2.500958040356636 1.2515618242323396 5.9151413999497873 1.6218358799815176 1.4712911956012247
leaf_count=47 21 23 63 38 33 19 23 31 26 31 35 25 39 24 25 18 19 31 28 28 27 19 34 18 40 30 14 150 24 17
internal_value=0 0.00254289 0.00597282 0.0412443 -0.00465861 -0.0133572 0.0468178 -0.00575707 -0.00430509 -0.0107568 0.0301378 -0.0256262 0.0211358 -0.014691 -0.0408984 -0.00957769 -0.0356107 -0.000361057 -0.0692359 -0.0239577 0.0693901 0.0232969 -0.0333495 -0.00686497 -0.0740443 -0.0730753 -0.0712812 0.0917813 -0.0947053 -0.105325
internal_weight=0 80.316 77.5777 17.9675 59.6102 50.9932 8.61698 9.3094 44.3743 40.831 10.8876 29.9434 7.37176 5.43136 22.5717 10.7215 6.71528 4.71854 11.8502 5.25599 7.84956 3.91181 8.25582 4.23387 6.61892 4.3449 3.03802 8.65811 4.89361 6.59419
internal_count=1000 979 932 292 640 566 74 104 483 457 126 331 88 63 243 107 81 62 136 56 81 46 82 41 83 48 45 188 64 80
is_linear=0
shrinkage=0.1
Tree=33
num_leaves=31
num_cat=0
split_feature=6 0 15 8 10 11 8 7 1 4 9 17 22 8 3 9 23 4 1 5 15 6 3 9 28 26 15 10 29 4
split_gain=1.99348 3.69945 3.43263 2.8833 2.39382 2.47127 2.11126 2.0819 1.79659 1.68928 4.19634 2.43661 2.17006 3.25037 2.10135 1.6853 1.44173 1.43402 1.24439 1.15854 1.0963 1.01426 0.884138 0.765431 0.726195 0.525422 0.442757 0.228505 0.189364 0.154436
threshold=2.090929508209229 0.44619986414909368 -0.8999110460281371 -1.1515020728111265 4.2750437259674081 4.8875966072082528 0.11012277007102968 -3.689206480979919 -3.7169742584228511 -0.44345408678054804 0.19106149673461917 1.1505017876625063 0.26117700338363653 1.701290547847748 2.2420258522033696 2.3375082015991215 -1.1823470592498777 -1.6112870573997495 -2.1899367570877071 -2.1659255027770992 0.040439540520310409 -1.4289640188217161 0.31626346707344061 -1.4426879882812498 0.74330320954322826 0.84513708949089061 0.84915056824684154 -1.3531174063682554 -0.67078268527984608 1.5504462718963625
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=3 -2 -3 7 5 8 -4 -1 -5 10 12 14 21 23 16 25 -11 -12 -18 -7 -15 -10 -23 -14 26 27 28 -9 -20 -6
right_child=1 2 6 4 29 19 -8 15 9 11 17 -13 13 20 -16 -17 18 -19 24 -21 -22 22 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.029311634796350003 -0.10654934190523693 -0.1165428939424677 0.10479334013801661 0.074658496689321868 -0.070445513901809129 0.009692363353183164 -0.039253157083884709 0.072192468208085656 -0.00059696672538711235 0.025097617650323478 0.10766676630504293 0.045919766309388987 0.028770019260129245 -0.1119961886236305 0.03381799240493074 -0.0064731620762225293 -0.01164788750431572 0.01509263436503344 -0.078525213885856468 0.095981210357101554 0.00064488440225842966 -0.11301400615964347 -0.028648074321905699 0.1096776328204196 -0.036694110082455454 0.03500947062043596 -0.048079472365497027 0.11776011240292893 -0.11757884602369797 -0.1188202172924686
leaf_weight=2.5580373536795369 4.7159171029925329 1.5580497384071348 2.9772485960274935 2.571269977837793 1.2441340070217841 2.6868332773447037 1.5457914117723701 1.4399232883006323 2.0800331793725428 1.777063868939879 4.3265642337501049 3.4661085382103911 2.3484267797321081 1.8268208131194117 2.8347665406763545 2.1039502173662186 2.8524915575981167 2.7285870425403109 1.5879505556076772 3.6968635730445376 1.6394692081958053 3.4660129044204968 1.9360418692231176 2.328871848061679 2.3743688315153113 1.2148311994969843 1.4234411641955373 4.6682835603132835 5.6920327115804072 1.4055007956922052
leaf_count=30 65 12 55 28 22 26 18 30 26 19 46 33 21 22 31 26 30 36 23 31 20 41 27 21 25 31 22 94 68 21
internal_value=0 -0.0400795 0.0114682 0.00616212 -0.00331972 0.00126326 0.0555641 0.0506983 -0.00662511 -0.0113019 0.00785002 -0.0310392 -0.0210528 0.0146679 -0.0454252 0.0724092 -0.0597266 0.0718636 -0.0705474 0.0596631 -0.0587198 -0.0599316 -0.0827781 0.0690547 -0.0857139 0.0950725 -0.0990869 0.107018 -0.10906 -0.0961059
internal_weight=0 10.797 6.08109 68.2787 56.2937 53.644 4.52304 11.985 47.2603 44.6891 22.6808 22.0082 15.6257 8.14359 18.5421 9.42699 15.7073 7.05515 13.9303 6.3837 3.46629 7.48209 5.40205 4.6773 11.0778 7.32304 8.70342 6.10821 7.27998 2.64963
internal_count=1000 150 85 850 639 596 73 211 539 511 260 251 178 84 218 181 187 82 168 57 42 94 68 42 138 155 113 124 91 43
is_linear=0
shrinkage=0.1
Tree=34
num_leaves=31
num_cat=0
split_feature=1 0 8 7 13 3 8 5 13 9 4 11 2 6 26 14 21 13 25 10 12 30 11 4 5 29 7 3 16 1
split_gain=1.9168 2.21045 2.41404 3.0026 1.94242 1.92721 2.23514 2.40801 1.69991 1.96772 1.93488 2.79849 1.61742 1.97029 1.67822 1.52284 1.48158 1.42378 1.60841 1.40312 1.36568 1.16496 1.05016 2.13352 1.10891 0.997295 0.494889 0.48285 0.479777 0.456111
threshold=4.1692261695861825 -3.1315989494323726 0.62742966413497936 -0.86519092321395863 1.4291877746582033 1.5328662991523745 -0.5797320008277892 0.76440930366516124 -0.45099973678588862 0.31265056133270269 0.61790910363197338 -0.70885902643203724 3.0034747123718266 0.77008655667305004 -1.0634013414382932 0.23209527134895327 1.1839795112609866 0.87251433730125438 -0.61387494206428517 -1.5083637237548826 0.45293231308460241 0.87502706050872814 -2.1287142038345332 -0.44345408678054804 -0.87853625416755665 -0.073632668703794465 -0.17649734020233152 -0.16505199670791623 -0.2259319871664047 -0.35371389985084528
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 -1 3 4 8 6 26 27 9 -3 11 29 13 14 -4 -11 17 18 -16 -15 -13 22 -20 -24 -25 -12 -5 -8 -27 -10
right_child=-2 2 12 5 -6 -7 7 -9 10 15 25 20 -14 19 16 -17 -18 -19 21 -21 -22 -23 23 24 -26 28 -28 -29 -30 -31
leaf_value=-0.091515854735354812 -0.098077597609280406 -0.04865642681674575 -0.094628987883676852 0.031719079596177931 0.082820609710801074 0.11042091029515828 -0.094469393883453601 0.074543241348465863 -0.024592059953682195 0.10704496433010412 -0.02337228220459521 0.10813479330972744 0.049704736515668611 0.019231935692900128 0.049473230981958244 0.0012471296681033884 -0.075380555831353752 0.062074342012911482 -0.069729592650713504 -0.097659583831142774 -0.0078419601740398782 -0.10009408442020214 0.079772777738573497 -0.094444244670164434 0.02218646007257322 -0.061413125465631768 0.10251688434594096 -0.019513191862175358 -0.13142256182483278 -0.097038936088107541
leaf_weight=2.4237786745652548 1.9480738416314114 2.3205924015492263 2.0732425563037493 1.3742006197571708 2.5766931101679793 4.6724003721028557 1.9303600974380972 2.0979491472244263 1.7365668788552286 3.1288916319608688 2.9434658568352461 2.471422208473081 3.4365271292626849 1.3020944260060785 3.8141716457903501 2.4072034582495689 2.3122254833579055 4.017089642584323 2.1996363773942091 4.8588529862463465 1.7233101110905407 1.7858023680746562 2.5609578061848977 1.8102057818323471 1.4831305928527927 1.7660442870110276 3.5072677731513977 1.5490460805594946 2.1961365751922131 1.7395249418914351
leaf_count=47 21 30 28 21 23 111 23 28 23 31 32 37 30 12 38 29 25 35 29 76 20 21 22 27 15 20 89 17 20 20
internal_value=0 0.00240985 0.00558075 0.0218641 0.000590777 0.0570261 0.0331723 -0.0100763 -0.00885422 0.0286409 -0.029064 0.00550465 -0.0150686 -0.0229571 -0.00899145 0.061042 -0.000106621 0.00974287 -0.00565347 -0.072955 0.0604884 -0.0270222 -0.01082 0.0113142 -0.0419204 -0.067463 0.0825863 -0.0610986 -0.100218 -0.0608463
internal_weight=0 74.2188 71.795 40.1411 25.0099 15.1312 10.4588 5.57736 22.4332 7.85669 14.5765 7.67082 31.6539 28.2174 22.0565 5.5361 19.9832 17.671 13.6539 6.16095 4.19473 9.83973 8.05393 5.85429 3.29334 6.90565 4.88147 3.47941 3.96218 3.47609
internal_count=1000 979 932 574 285 289 178 68 262 90 172 100 358 328 240 60 212 187 152 88 57 114 93 64 42 72 110 40 40 43
is_linear=0
shrinkage=0.1
Tree=35
num_leaves=31
num_cat=0
split_feature=3 8 4 11 2 9 9 6 30 4 6 7 23 6 23 10 3 3 26 9 8 31 11 6 8 13 0 27 15 22
split_gain=1.90706 2.01671 3.09355 2.48163 2.29895 2.24932 2.10222 2.83775 2.54511 2.1898 1.92346 1.72085 1.73671 2.17558 1.69311 1.77402 1.58648 1.5785 1.56864 1.26894 1.22321 1.19037 1.04313 0.925242 0.882426 0.837422 0.828022 0.790375 0.662162 0.466266
threshold=-1.9173253774642942 2.4709751605987553 -3.4333997964859004 -0.97970148921012867 1.1303530335426333 -0.50164487957954396 -1.9159966707229612 2.2315721511840825 1.2294318079948428 1.7565276622772219 -0.59109562635421742 -1.3801147341728208 0.57123467326164257 -0.45363011956214899 0.16183397918939593 -1.5789780020713804 1.4957433342933657 -0.9574308693408965 -0.19718474894762036 -2.6924784183502193 0.69922864437103283 -0.51559561491012562 0.91672056913375866 -0.68222475051879872 0.071703881025314345 0.40194204449653631 -0.49639803171157831 0.40907128155231481 -0.8999110460281371 0.14942352473735812
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=14 2 19 4 5 29 22 8 9 11 -11 23 13 -13 15 25 -12 -7 -3 -2 -19 -20 -5 -8 -15 -1 -6 -24 -9 -4
right_child=1 18 3 6 26 17 7 28 -10 10 16 12 -14 24 -16 -17 -18 20 21 -21 -22 -23 27 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.013722844711203778 -0.020209907650366996 -0.10933330323697632 -0.094437408235584705 -0.0043992649346422203 -0.12036441201513015 -0.064088934300855008 0.016409584432236213 -0.12462184436443165 -0.07474070725248512 -0.077043324955500828 0.074394104493575738 0.070429548130881023 0.083296940402275726 0.001617760820290963 -0.092966676702791573 -0.083874393895083735 -0.036680334710902839 0.09035609616602959 0.033601438445724545 0.10719169538300928 -0.0084629858118747247 -0.061553941787633105 0.1218487358851696 0.10483248513222683 -0.10077363432202038 0.087613187945893289 -0.033689655829755312 0.035744135666855255 -0.037164112691567878 -0.021371739685260474
leaf_weight=1.5722512751817705 0.95004094205796707 2.3973473785445094 2.258816702291369 1.3936009444296278 2.8987642154097557 1.3977733347564925 1.4858520925045011 1.5766020603477957 2.5976212862879029 2.9521070290356874 2.8377512469887716 2.0157192517071945 2.8900117073208094 1.400445157662052 3.8917094245553017 1.8426991924643497 2.3514292594045401 3.5033153612166643 2.7575408741831779 4.4145853053778401 1.9497731551527975 2.5125238429754964 3.2682410422712564 5.813293579965829 2.1095733102411032 1.6942512392997739 1.7783792568370689 1.5821267962455747 1.9199101226404298 1.423980865627527
leaf_count=22 10 39 30 17 42 24 18 13 35 38 25 28 33 27 55 29 24 66 26 81 26 37 43 61 24 18 35 21 30 23
internal_value=0 0.00586712 0.0123687 0.00482757 -0.0290654 -0.00316002 0.0190712 0.00806728 0.019258 0.0294933 -0.0125999 0.0513001 0.020482 -0.0123708 -0.0432725 -0.00542008 0.0240617 0.0307207 -0.0422709 0.0846297 0.0550229 -0.0117643 0.0718537 0.0868327 -0.059921 0.0388376 -0.0874083 0.0937625 -0.0765994 -0.066186
internal_weight=0 64.4371 56.7697 51.4051 15.2108 10.5337 36.1943 29.9503 26.4538 23.8562 8.14129 15.7149 8.41575 5.52574 9.00091 5.1092 5.18918 6.85086 7.66741 5.36463 5.45309 5.27006 6.24397 7.29915 3.51002 3.2665 4.67714 4.85037 3.49651 3.6828
internal_count=1000 876 774 683 246 169 437 356 313 278 87 191 112 79 124 69 49 116 102 91 92 63 81 79 51 40 77 64 43 53
is_linear=0
shrinkage=0.1
Tree=36
num_leaves=31
num_cat=0
split_feature=13 1 0 8 7 13 6 8 3 17 31 3 4 9 11 17 24 9 7 20 21 29 27 11 4 13 22 10 8 0
split_gain=1.68247 2.33494 2.14333 1.76531 2.76868 2.21798 2.17383 2.32163 2.04419 1.51841 2.45095 1.74325 1.91785 1.68409 1.92969 1.61659 1.44551 1.40017 1.37415 1.21567 1.20768 1.1646 1.09462 1.05653 0.972171 0.932699 0.871931 0.694889 0.544738 0.315358
threshold=-1.2598152160644529 4.1692261695861825 -3.1315989494323726 2.9477462768554692 0.34028041362762457 1.3883623480796816 -1.1788545250892637 0.72140043973922741 -1.9173253774642942 0.75022095441818248 1.3007090091705324 2.4212653636932377 -0.38929480314254755 -0.39009538292884821 -0.13035058230161664 -0.53055807948112477 1.0000000180025095e-35 1.4011751413345339 -1.6642212867736814 -0.45171776413917536 0.034177251160144813 0.0088136638514697569 0.23536460846662524 2.9679244756698613 -1.5022985935211179 -0.10182148218154906 -0.39222112298011774 -0.62594765424728382 0.14835035800933841 -0.34969562292098993
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=20 2 -2 4 5 8 -6 19 24 11 28 12 13 14 27 25 -9 23 -7 -8 -1 -22 -5 -17 -4 -14 -13 -10 -11 -15
right_child=1 -3 3 22 6 18 7 16 9 10 -12 26 15 29 -16 17 -18 -19 -20 -21 21 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.08763952203152775 -0.10812950574693742 -0.12788402506877777 -0.0044199921724342164 -0.11002500552497013 0.11335782823056216 0.091080810237702703 0.001995613034971172 -0.11205152295019197 -0.0021837581615992397 -0.048327277647770263 0.060306120731685314 -0.010243540722944915 -0.11304792906190396 0.045793441144258096 0.063440774038661524 -0.083373274287676782 0.012593904203915044 0.045377112845350655 -0.0072174636554795868 0.10824956532606222 0.068173703756496992 -0.041643993598361496 -0.011066300350886508 0.021622181468053371 -0.10258237476881156 -0.02920763369415464 0.096730329124082318 -0.094740449932983792 -0.11363691189912684 0.10978178989669002
leaf_weight=3.4967069784179321 1.8600791143253435 1.5197678580880154 1.4076077304780481 2.1918310672044745 3.5855409260839215 3.4856420457363129 1.7169054728001345 1.5283866524696352 1.4630701187998059 2.4778521824628106 1.5858978759497402 1.0498005617409947 2.7662662379443645 1.218195414170623 2.4746550805866718 2.3889197893440932 2.3779947366565466 3.5050492696464062 2.4022491322830311 2.8880551271140567 1.7276070415973666 2.1895917346701026 2.2810418382287025 1.6004708558320997 3.5619751457124869 2.5500965584069428 2.7788669792935243 1.8204073794186113 2.635500842705369 2.0943200057372415
leaf_count=42 42 18 16 32 73 30 30 21 19 30 16 17 29 16 29 33 28 36 35 75 26 28 26 16 60 30 40 30 39 38
internal_value=0 -0.00543217 -0.00241582 0.000870392 0.00575251 -0.006072 0.0480451 0.0205311 -0.015058 -0.00590063 -0.0483037 0.00514829 -0.00574386 0.0294396 -0.00324188 -0.0306554 -0.0361741 -0.000736196 0.0509753 0.068634 0.0449215 0.00678904 -0.0595588 -0.041251 -0.0747784 -0.0728323 0.0673987 -0.0534985 -0.0819889 0.0862497
internal_weight=0 63.2164 61.6967 59.8366 55.3637 43.2668 12.0969 8.51134 37.379 32.4094 6.69925 25.7101 21.8815 9.07065 5.75813 12.8108 3.90638 7.49444 5.88789 4.60496 7.41391 3.9172 4.47287 3.98939 4.96958 5.31636 3.82867 3.28348 5.11335 3.31252
internal_count=1000 904 886 844 786 559 227 154 494 418 85 333 276 132 78 144 49 85 65 105 96 54 58 49 76 59 57 49 69 54
is_linear=0
shrinkage=0.1
Tree=37
num_leaves=31
num_cat=0
split_feature=8 1 10 2 6 8 4 9 30 3 9 12 28 19 3 0 20 3 5 2 3 9 14 1 12 21 30 21 1 28
split_gain=1.5427 1.99274 1.93302 1.90644 2.7912 2.04429 2.13545 1.88489 1.99672 2.49246 1.63129 2.07452 1.60124 1.42708 1.41397 1.41447 1.36972 1.73912 1.32378 2.21153 1.17342 1.14806 1.1209 1.08124 0.917969 0.910457 0.855787 0.80714 0.744997 0.493074
threshold=-2.4568742513656612 3.704098224639893 4.8179185390472421 3.0348654985427861 2.0550829172134404 1.5303894281387331 -0.1948781609535217 -1.6665836572647093 0.38807439804077154 -0.98315218091011036 0.12808340787887576 0.079420387744903578 -0.27788183093070978 0.25574015080928808 1.7003743648529055 -0.78594365715980519 0.63956779241561901 -0.9943212866783141 -0.25183287262916559 0.039535939693450935 -0.81044566631317128 -1.7018492817878721 -0.67656695842742909 -1.3355131745338438 0.018321231938898567 0.31665398180484777 -0.68899607658386219 -0.024068400263786312 0.56481224298477184 0.65189185738563549
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=14 2 3 4 5 6 10 -8 9 -9 11 22 -11 23 15 -1 17 -5 19 -15 -13 -22 -2 -7 -17 -24 -14 -10 -4 -12
right_child=1 -3 28 16 -6 13 7 8 27 12 29 20 26 18 -16 24 -18 -19 -20 -21 21 -23 25 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.099109164741745248 0.1284570488371547 -0.097991815298368978 -0.10563658825387007 -0.058109535507103231 -0.082200190923889124 -0.0027717696277348176 0.081561060838665211 -0.089506169506829256 -0.031472489539085532 -0.030002849395967247 0.10285535824218046 -0.082380081555884249 0.010901993050254825 0.051960678049686918 0.10851330631770364 0.11028896307911971 0.12801712575292717 0.079505181502764297 0.061667072671800882 -0.10285862210798619 -0.054918059902694029 0.075906846982271403 -0.020816694998772894 -0.086053880028266846 -0.013875610790179869 0.082605095892690439 0.093851240014381998 -0.11626043485888969 -0.014360509753926567 0.039166502517888321
leaf_weight=0.9052511136978868 1.7448244392871801 2.2087648548185816 2.3242425639182325 1.5436387434601786 4.0710279494523993 2.2966701965779066 2.0984227340668467 2.2228493373841038 2.1694798599928617 2.887221733108162 4.3142980234697452 2.1755066625773907 1.9967267001047719 1.7022542357444765 2.3790571074932814 1.1819119127467268 1.8827211419120429 2.2670079711824647 2.5439618453383446 2.0146578624844551 1.2326312419027092 1.4716317960992453 1.7360467696562412 4.8528074976056796 1.1999538308009503 1.6700595617294309 3.2982991505414319 2.3270129896700373 1.453371040523052 1.6924448097124694
leaf_count=13 18 27 42 24 67 34 26 37 28 35 83 22 20 14 62 27 17 24 24 34 20 26 30 82 26 23 37 28 21 29
internal_value=0 -0.00470966 -0.00127511 0.00337832 -0.00280403 0.00415483 0.0175209 -0.00717263 -0.019668 0.00439536 0.0436966 0.0190163 0.0299053 -0.0287736 0.0497942 0.00729608 0.0582361 0.0237594 0.00608606 -0.0319552 -0.0277073 0.0162755 0.0632802 -0.0593007 0.0477364 0.0298924 0.0625715 -0.0753517 -0.0705197 0.0849105
internal_weight=0 62.1986 59.9898 56.2122 50.5188 46.4478 33.0375 17 14.9016 10.4051 16.0374 10.0307 8.18225 13.4104 5.66617 3.28712 5.69337 3.81065 6.26087 3.71691 4.87977 2.70426 5.15093 7.14948 2.38187 3.40611 5.29503 4.49649 3.77761 6.00674
internal_count=1000 872 845 782 717 650 462 211 185 129 251 139 92 188 128 66 65 48 72 48 68 46 71 116 53 53 57 56 63 112
is_linear=0
shrinkage=0.1
Tree=38
num_leaves=31
num_cat=0
split_feature=13 1 0 8 7 4 3 30 27 22 3 9 2 16 25 2 5 19 9 4 18 8 6 19 11 30 3 27 29 8
split_gain=1.37742 1.79953 1.81635 1.5789 2.41294 1.36158 1.62003 1.13254 1.12128 1.63086 1.20063 1.09093 1.03404 1.25648 1.68074 1.4993 2.25611 1.25203 1.54091 1.88163 0.999464 0.929165 1.28507 0.914617 0.984544 0.909855 0.853064 0.615563 0.613342 0.2563
threshold=-1.2598152160644529 4.1692261695861825 -3.1315989494323726 0.62742966413497936 -0.20904585719108579 3.1555353403091435 2.6009738445281987 -1.2419461011886594 0.31288018822669988 0.76032915711402904 -1.9173253774642942 -0.2785354852676391 -2.1278877258300777 0.25312055647373205 0.97147566080093395 3.0034747123718266 -2.0176103115081783 0.68507990241050731 0.97868943214416515 -1.8213946819305418 0.009796811733394863 -0.52487343549728382 1.3110203742980959 0.16267675906419757 0.24271030724048617 -0.28692136704921717 -0.49797442555427546 -0.58443316817283619 0.11635308712720872 -0.82658091187477101
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=11 2 -2 4 5 6 8 -5 9 23 -10 -1 -9 14 15 16 -14 18 19 -18 -12 -6 26 29 -25 -15 -23 -11 -13 -4
right_child=1 -3 3 7 21 -7 -8 12 10 27 20 28 13 25 -16 -17 17 -19 -20 -21 -22 22 -24 24 -26 -27 -28 -29 -30 -31
leaf_value=-0.023462554640897366 -0.10699804036056014 -0.1178964568135133 -0.05511555660771892 -0.088134650740420789 0.098110284746367085 -0.076397786276596907 0.084262318697305286 0.074111212586162684 -0.062188299647647431 0.072099528770015797 0.0028120205355044599 0.096955099323459415 -0.10551166491888186 0.0019789104580798236 -0.082576279102188815 0.09808334461029361 0.048128802567999764 0.070642119914300472 0.071256096420854453 -0.089113224261146415 0.094362314710357553 -0.012573232697070742 -0.051249820420496778 -0.056682862337726583 0.048564794783135724 -0.076529974690862312 0.10358748374492602 -0.010848451752638023 0.026904863150583016 -0.1047991680848416
leaf_weight=1.8622148670256136 1.6136616310104717 1.3823975026607502 2.0958228539675523 2.1000747792422798 3.8689672620967031 2.167483150959014 2.3910445049405089 1.3316118866205275 1.1074621491134164 1.9333429597318175 2.1633529132232061 2.8319404581561676 1.7231521159410519 2.1461413130164146 2.1286591906100503 1.796665247529744 1.566128365695479 3.1225909646600485 1.7742460053414104 2.7586633767932556 2.6571020716801286 0.91955195087939534 1.32689718529582 2.0455251652747397 1.5717732235789297 4.7285951003432265 2.0232253614813107 1.6652871835976837 2.2374631548300385 2.0577234849333803
leaf_count=33 42 18 28 30 143 23 39 10 19 20 34 33 35 33 25 14 18 33 18 40 33 15 13 31 21 76 42 23 30 28
internal_value=0 -0.00516619 -0.00242183 0.000636851 0.0161356 -0.00117221 0.00710931 -0.017828 -0.00355566 -0.02194 0.0317046 0.0419927 -0.0114297 -0.016668 -0.000323581 0.013418 -0.000480399 0.0191457 -0.00721944 -0.039414 0.0532758 0.062615 0.030451 -0.0477134 -0.0109511 -0.0520212 0.0672899 0.0337149 0.0660373 -0.0797295
internal_weight=0 58.1671 56.7848 55.1711 29.9946 21.8559 19.6884 25.1765 17.2974 11.3695 5.92792 6.93162 23.0765 21.7448 14.8701 12.7414 10.9448 9.22163 6.09904 4.32479 4.82045 8.13864 4.26967 7.77084 3.6173 6.87474 2.94278 3.59863 5.0694 4.15355
internal_count=1000 904 886 844 512 299 276 332 237 151 86 96 302 292 183 158 144 109 76 58 67 213 70 108 52 109 57 43 63 56
is_linear=0
shrinkage=0.1
Tree=39
num_leaves=31
num_cat=0
split_feature=3 2 6 8 2 3 11 4 30 23 10 5 30 30 28 12 19 12 13 20 15 4 11 8 31 1 7 9 7 11
split_gain=1.23429 1.42246 1.60061 1.53313 2.16457 1.35334 2.1318 1.59076 1.36907 1.31874 1.37771 1.31532 1.67996 1.77445 1.13604 1.11382 1.06975 1.16318 1.04744 1.04034 0.878914 0.870451 0.849188 0.842901 0.804078 0.929988 0.682372 0.920979 0.551401 0.495999
threshold=-1.9173253774642942 4.188784360885621 2.2315721511840825 0.84256586432456981 3.4516642093658452 1.5328662991523745 2.7736781835556035 -3.4333997964859004 -0.2332557141780853 0.16183397918939593 -1.5789780020713804 -2.1957887411117549 -1.2419461011886594 0.55339282751083385 -0.91644141077995289 -0.89584058523178089 0.82704880833625805 0.67080923914909374 0.55809754133224498 -1.1364628076553343 -0.3727851659059524 1.7364857196807864 -0.43148031830787653 1.3867889642715456 0.38249349594116216 -0.20998486876487729 0.4275700449943543 2.193432211875916 -1.1648204326629636 0.49007196724414831
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=9 2 3 4 5 6 7 -2 14 10 29 15 -13 16 -9 -5 17 22 28 -10 -15 -7 23 -14 25 -16 27 -21 -4 -1
right_child=1 -3 18 11 -6 21 -8 8 19 -11 -12 12 13 20 24 -17 -18 -19 -20 26 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.0030086245689277848 0.085275220549334854 0.083710176725187713 -0.11705783360314968 0.020089548454868607 -0.073826495484712137 0.087688547796910293 0.08084907898463578 -0.062220546421413017 0.047608171413430926 -0.086197841846712905 -0.080255560427582626 -0.11474915517916376 0.044474798787780122 0.018060176684975979 0.087189947244240784 -0.10662689545317797 0.079926235739169085 -0.059612209942153216 0.027045473274262394 -0.097161988745415262 -0.086723774356071634 0.012429205279843213 0.059053278857242958 -0.062758788894871811 0.11076122681222267 -0.036864475704776453 0.0094741259788656967 -0.0044941331907458566 -0.038852262192116134 0.080407985588114461
leaf_weight=1.4244034346193077 1.8340979674831057 2.206948426552116 1.9073003344237802 0.96961598657071624 2.1661289297044268 5.3464033082127562 3.9606026094406834 1.1254151836037634 1.1417509559541965 3.2448534909635782 1.4771261513233174 1.2294935155659974 1.3865315131843106 1.1718427911400797 1.0672570103779437 2.4373380169272449 2.7958496734499922 1.9323869179934261 1.2204524409025905 3.8686447162181139 2.5260441694408646 2.1567905228585005 3.8728133682161503 1.5552134960889814 1.4533010069280861 1.3930967934429674 1.2930003199726341 1.483844321221113 1.7097141863778227 1.4268512856215236
leaf_count=26 37 34 21 13 25 134 43 16 14 55 29 14 18 16 18 60 26 29 31 51 43 29 36 30 18 25 34 24 37 14
internal_value=0 0.00504815 0.00177293 0.00727995 0.0222344 0.0301994 0.0157516 -0.00183489 -0.0142912 -0.0380025 -0.00187203 -0.0140041 -0.00230448 0.00676664 0.0263231 -0.0705635 0.0260797 0.00886834 -0.0530614 -0.0405724 -0.0535182 0.0660553 0.0282872 -0.0122164 0.0517848 0.0169481 -0.0557226 -0.0714721 -0.0800911 0.0387355
internal_weight=0 55.2119 53.0049 48.1675 28.2903 26.1242 18.621 14.6604 12.8263 7.57323 4.32838 19.8771 16.4702 15.2407 5.03907 3.40695 11.5428 8.74695 4.83747 7.78724 3.69789 7.50319 6.81456 2.94175 3.91365 2.46035 6.64549 5.35249 3.61701 2.85125
internal_count=1000 876 842 753 468 443 280 237 200 124 69 285 212 198 77 73 139 113 89 123 59 163 84 48 61 43 109 75 58 40
is_linear=0
shrinkage=0.1
Tree=40
num_leaves=31
num_cat=0
split_feature=0 8 2 1 6 11 4 22 4 6 13 5 11 5 18 3 1 31 1 6 28 29 8 20 3 18 12 4 3 4
split_gain=1.20182 2.03383 1.90995 1.68508 1.92889 2.13362 1.67942 1.59175 1.6941 1.56317 1.54824 1.8315 1.31249 1.69012 1.31081 1.24976 1.21644 1.08995 1.04577 0.966923 1.01667 0.907286 0.76985 0.721162 0.634353 0.974244 0.224232 0.220417 0.156752 0.124349
threshold=-0.29508113861083979 -0.82658091187477101 -0.49497580528259272 -0.33602377772331232 -1.054078161716461 4.6643924713134775 0.69948077201843273 0.76032915711402904 2.0288132429122929 -0.44430741667747492 0.84571009874343883 -0.25183287262916559 -0.056552402675151818 -2.5214469432830806 -0.59103864431381214 -0.053450405597686761 0.86920762062072765 -1.6753969192504881 0.6997278928756715 -0.59109562635421742 -0.50307175517082203 -0.047192480415105813 0.071703881025314345 0.034880844876170165 0.37997007369995123 0.031244269572198394 1.3434526324272158 1.5504462718963625 1.2886238098144533 -1.4345278739929197
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=2 6 22 7 14 17 26 8 9 -3 11 -11 24 27 -5 -7 -8 -6 21 -15 -21 -19 28 -16 29 -26 -2 -14 -1 -4
right_child=1 3 12 4 5 15 16 -9 -10 10 -12 -13 13 19 23 -17 -18 18 -20 20 -22 -23 -24 -25 25 -27 -28 -29 -30 -31
leaf_value=0.054577271195799741 0.10485100180808904 -0.078125350528555793 -0.05998400199731127 -0.035865468765675722 0.025759827678277631 -0.053183282732058759 0.057912915645312035 0.086832733726449715 0.091829881336931429 -0.066575281368772429 0.082758303468390151 0.059273230817973026 -0.0457343315788169 -0.046475772203833129 0.012866672359935409 0.060658415763545519 -0.055965173185119371 0.028039188163628395 -0.1081298720922773 -0.016434940554604271 0.080721106477513932 -0.08864573055987518 -0.0068362988479838561 0.10429553315609266 0.0089802962882399393 -0.10035216992475585 0.033555107292501933 -0.11029117489344711 0.11149998086678309 -0.10529269555538714
leaf_weight=0.85217270068824591 3.8452951833605749 2.5265795821323973 0.76026600226759944 2.0422408692538738 1.0565537139773331 1.5550388470292089 2.0376345477998261 3.1462902612984172 1.941201839596032 2.9969258420169336 2.2172956969588995 1.8829800020903364 0.87985815107822452 1.7152472995221613 1.5170915201306336 2.5385410115122786 1.7381913773715494 1.3103691935539248 5.2798633947968483 1.7968621831387279 2.6886820616200513 1.3558845762163452 1.5777071351185439 2.000088413245976 1.9713657693937423 1.3894715905189512 0.49829760566353787 1.3258583936840294 1.1190541647374628 2.9800509121269014
leaf_count=18 151 34 12 28 11 25 27 35 23 35 25 23 20 28 20 30 22 23 79 19 31 24 24 22 35 32 19 30 23 72
internal_value=0 0.00938383 -0.0209535 -0.00153834 -0.0214943 -0.0424445 0.0542687 0.0237682 0.00661127 -0.0105781 0.0134685 -0.0180149 -0.0360983 -0.00936008 0.027858 0.0174131 0.00548944 -0.0696622 -0.08235 0.0173824 0.0418014 -0.0312992 0.0452243 0.0648588 -0.0677516 -0.036221 0.0966719 -0.0845395 0.086892 -0.0960831
internal_weight=0 41.4864 19.0566 33.3669 18.6557 13.0963 8.11942 14.7113 11.565 9.62378 7.0972 4.87991 15.5077 8.40651 5.55942 4.09358 3.77583 9.00267 7.94612 6.20079 4.48554 2.66625 3.54893 3.51718 7.10115 3.36084 4.34359 2.20572 1.97123 3.74032
internal_count=1000 656 344 437 262 192 219 175 140 117 83 58 279 128 70 55 49 137 126 78 50 47 65 42 151 67 170 50 41 84
is_linear=0
shrinkage=0.1
Tree=41
num_leaves=31
num_cat=0
split_feature=22 4 4 9 11 9 6 5 1 13 8 30 3 16 28 7 27 13 8 30 15 31 14 3 8 21 4 10 9 5
split_gain=1.0924 1.37256 1.31385 2.7415 1.87623 2.08519 1.56976 1.26345 1.4826 1.27613 1.2646 1.27363 1.2325 1.2077 1.10277 1.01511 1.04834 1.16622 1.10717 1.18832 0.954239 1.36657 0.878716 0.857174 0.822412 0.684123 0.642391 0.594322 0.591547 0.470641
threshold=0.88484010100364696 1.4679937958717348 -3.1178826093673702 -1.1903105974197385 -0.056552402675151818 -1.8577739000320432 2.2315721511840825 -0.15890216827392575 1.4979884624481203 1.279394865036011 0.93145474791526806 1.2621341943740847 2.2192145586013798 0.64496758580207836 0.1477523744106293 -1.0754944086074827 -1.0113512277603147 1.0001536607742312 1.3702826499938967 0.38249950110912329 0.18318539112806323 0.49908749759197241 0.13633336126804355 -0.37849885225296015 1.5303894281387331 0.42853437364101415 -1.5022985935211179 1.234558701515198 1.6468444466590884 -1.0000000180025095e-35
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=2 22 3 23 12 25 7 9 -9 10 11 20 15 -12 -3 28 -17 18 19 29 -7 -22 24 -1 27 -6 -24 -2 -4 -18
right_child=1 14 4 -5 5 6 -8 8 -10 -11 13 -13 -14 -15 -16 16 17 -19 -20 -21 21 -23 26 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.025576806506602726 0.076627257824640638 -0.088112357647712941 -0.097958382046783796 0.10515403474837547 0.097296151920377882 0.062191894997988752 -0.087923250251060944 0.09336641889039915 -0.030058319336285434 0.049632226219447684 -0.10017911705276106 -0.083621756842135492 0.055474828027708006 0.012769142208168389 0.028960394167067183 0.065296220787446166 -0.00027455440404420455 -0.11792154836442298 -0.1061100856016586 -0.069530169835275782 0.043673118293344373 -0.089878174597100846 0.020957334844720132 -0.091887083117135115 -0.045440749969977927 0.01664652426023068 0.10129642676563566 -0.00910273955860357 -0.028425008757086524 0.092713330421430179
leaf_weight=1.1515145348384979 2.0517576951533543 1.6426789332181209 4.7042335439473382 2.8946562162600449 2.6554132765159002 2.7252397621050495 2.3215955486521116 2.6325945509597757 1.5440670419484375 2.5489497706294051 3.3958954932168126 1.5240200934931634 1.1855091778561462 1.3125797486864028 1.5770212393254039 1.0921810911968339 0.9260349180549371 1.4890859648585317 1.0399405490607021 1.1807381482794879 1.6308308560401203 1.4451395254582164 1.45107658021152 1.3490543607622383 1.5329787842929361 1.7416425682604311 3.1685931896790862 1.3346585165709255 1.6535598123446105 1.320383398793636
leaf_count=15 32 24 85 74 45 40 40 35 21 27 65 20 35 19 20 17 23 27 24 27 29 18 32 30 28 22 39 28 34 25
internal_value=0 0.0256859 -0.00742705 0.0389004 -0.0136649 0.00271104 -0.0103545 -0.000754816 0.0477376 -0.0146437 -0.0282584 -0.00226839 -0.0422582 -0.0686926 -0.0307697 -0.0509008 -0.0247663 -0.0412811 -0.0157333 0.0116907 0.0191037 -0.0190714 0.0447412 -0.0377948 0.0153295 0.0653514 0.0760613 0.0428392 -0.0798739 0.0543812
internal_weight=0 12.7588 45.4649 5.39523 40.0696 25.478 21.0809 18.7593 4.17666 14.5827 12.0337 7.32523 14.5917 4.70848 3.2197 13.4062 7.04836 5.95618 4.4671 3.42716 5.80121 3.07597 9.53906 2.50057 4.91939 4.39706 4.61967 3.38642 6.35779 2.24642
internal_count=1000 203 797 119 678 381 314 274 56 218 191 107 297 84 44 262 143 126 99 75 87 47 159 45 88 67 71 60 119 48
is_linear=0
shrinkage=0.1
Tree=42
num_leaves=31
num_cat=0
split_feature=4 9 9 11 26 2 6 8 14 5 1 17 10 3 3 6 1 10 16 11 29 3 22 17 21 4 26 18 5 7
split_gain=1.04226 2.68099 2.18703 1.8076 1.59403 1.56837 1.85815 1.62072 1.476 1.47031 2.15766 1.60625 1.52068 1.13509 1.11375 1.53452 1.26554 1.04796 0.995707 0.945789 0.73632 0.669235 0.597073 0.577342 0.482225 0.475245 0.438839 0.40725 0.799752 0.312462
threshold=-1.0175390839576719 -2.2834039926528926 -1.9159966707229612 0.91672056913375866 -0.82819029688835133 4.188784360885621 2.090929508209229 -1.8904331922531126 -0.57414984703063954 -0.87118589878082264 -2.1300677061080928 1.0725740790367129 -0.25624310970306391 -1.0000000180025095e-35 1.4348896741867068 -0.60408353805541981 0.3707870244979859 -0.14712433516979215 -0.77104368805885304 0.15818990021944049 -0.031061415560543534 1.4703543782234194 0.61048614978790294 0.68529933691024791 0.12501231953501704 0.35684651136398321 0.20450743287801745 -0.074927981942892061 -1.059522747993469 -1.0066578388214109
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 18 3 -2 8 6 7 21 -3 14 19 12 22 17 15 -9 -17 20 -1 -11 -6 -4 29 -10 -5 -16 -19 28 -15 -12
right_child=2 4 5 24 13 -7 -8 9 23 10 11 -13 -14 27 25 16 -18 26 -20 -21 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.033171451996477891 -0.048107987324232222 -0.13226958166950983 0.011054897699661802 0.11967885550722329 0.088008425676158297 0.070300618082235519 -0.099272464692484033 -0.10637028664104481 0.069926793242119811 0.010089848091580892 -0.11183130516973022 0.054083226534479924 0.032827156384888992 0.00023802085719376679 -0.03794307173038796 0.060863452070726624 -0.045051986575280062 0.012203406850580219 -0.082394221755679634 0.11300060887872497 0.00099952449757229923 0.10600052202625364 -0.011960754969705702 -0.039502728958207285 0.044215199103542137 -0.11256679035890292 -0.077118365539707107 0.10828099481206591 0.099563326727689916 -0.059090506864359862
leaf_weight=0.98657224792987097 1.3056303123012174 0.94461333565413985 1.7180110830813711 2.124091974459589 2.384852203540504 1.8218950051814307 3.0465174466371527 1.7360126767307518 1.0589414536952997 1.7369179408997362 2.6772396788001114 1.9039705023169515 1.7641972508281281 1.3973589572124185 1.1698473170399664 2.1082641072571366 2.4265962261706591 1.0799875911325219 3.0516606243327287 1.8381208665668962 1.6424489952623842 1.3072863435372708 1.2571823196485636 0.88511980138719071 1.4081611325964329 3.1551497420296073 1.1209144555032251 3.0884732739068514 1.9307397357188163 1.9353631483390934
leaf_count=17 22 13 29 28 38 25 57 47 27 27 43 24 33 28 27 26 39 19 63 22 33 30 18 19 22 50 21 77 41 35
internal_value=0 0.0184286 -0.0101817 0.052432 0.0372996 -0.0197669 -0.025277 -0.0168448 -0.0297229 -0.0256401 -0.00325457 -0.028089 -0.0485833 0.0526108 -0.0533434 -0.0264185 0.00418828 0.0221999 -0.0541606 0.0630018 0.0525237 0.0520826 -0.0730517 0.0201042 0.0895947 -0.0923822 -0.033288 0.082129 0.0578599 -0.0897022
internal_weight=0 19.5717 36.4405 4.83788 15.5334 31.6026 29.7807 26.7342 2.88867 23.7089 13.113 9.53795 7.63398 12.6448 10.5959 6.27087 4.53486 6.2282 4.03823 3.57504 4.0273 3.0253 5.86979 1.94406 3.53225 4.325 2.2009 6.41657 3.3281 4.6126
internal_count=1000 396 604 72 316 532 507 450 59 391 202 153 129 257 189 112 65 111 80 49 71 59 96 46 50 77 40 146 69 78
is_linear=0
shrinkage=0.1
Tree=43
num_leaves=31
num_cat=0
split_feature=18 8 4 3 4 16 30 6 7 16 6 12 0 16 31 18 0 1 9 2 3 22 26 3 23 26 9 29 11 7
split_gain=1.02202 1.1977 1.61313 1.63954 1.3726 1.75708 1.33121 1.29957 1.7957 1.69565 1.45169 1.21385 1.19889 1.02073 0.996881 0.980227 1.0088 1.21027 0.901985 0.891209 0.875844 0.862495 0.804326 0.639749 0.810784 0.493048 0.488521 0.420318 0.383077 0.462897
threshold=-1.5012420415878294 1.7825397253036501 -3.3526372909545894 -0.24213722348213193 0.72157731652259838 -0.63479617238044728 0.50428083539009105 1.7042843699455263 -2.1902713775634761 -0.21955417096614835 -0.59109562635421742 -1.2767904400825498 -0.10111981630325316 -1.1816231012344358 1.3007090091705324 0.56845283508300792 0.70602595806121837 1.2682092189788821 -2.6924784183502193 1.5823640227317812 0.66349661350250255 0.28407719731330877 -0.09652735665440558 0.78698328137397777 -0.11595648527145384 -0.32592111825942988 2.4191335439682011 -0.38712324202060694 0.32232411205768591 -0.80342191457748402
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=23 2 18 11 5 -3 7 8 10 -9 -5 -4 27 -10 15 16 25 19 -2 -18 -6 22 -14 24 -1 -13 28 -8 29 -15
right_child=1 4 3 6 20 -7 12 9 13 -11 -12 14 21 26 -16 -17 17 -19 -20 -21 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.076035273392340175 -0.029346830339664051 0.015126151064815642 0.038288415757061656 -0.09286645714115957 0.057476994803689209 -0.097622740348237952 -0.1308270332333856 0.06757970751083904 -0.025090429634429687 -0.08521365973793471 0.036859119363592631 -0.019262062370717831 0.031318784975676944 -0.0034136683262903107 0.027248206621510784 -0.10171286385600076 -0.014985562171147033 -0.059042639819376699 0.099514911545899337 0.10531887759949084 -0.048085591300430941 0.075078489930381076 -0.090726986526686301 0.097978973923031787 -0.040335736933423218 -0.10150334599897654 0.018544025956713232 -0.045764216072387204 0.10894914799734906 0.086524557523427872
leaf_weight=1.16102730575949 0.69316109456121955 2.0093783438205719 2.3440126609057179 1.4490803396329286 2.0374404108151798 4.4281815541908127 0.95369897689670358 1.1634300677105787 1.0464774314314063 1.9331872286275018 2.1314668059349016 1.0721022272482517 1.1330319200642409 0.79530518595129285 1.874290077947079 2.879524321295321 1.2437915913760664 2.0783718721941113 2.5105969910509875 1.2195141166448591 1.2795910732820628 1.3467125925235448 1.0316774053499096 1.5805867314338682 1.2361636720597735 2.2776253437623328 1.1207895837724207 1.4860441051423547 4.1663768477737904 2.0405295118689679
leaf_count=17 9 28 35 23 26 94 15 21 21 31 34 22 26 16 26 51 23 35 81 17 29 28 16 26 25 50 18 36 69 52
internal_value=0 -0.00408573 0.00357769 -0.00234934 -0.0355034 -0.06243 0.0151571 0.0303014 0.0444144 -0.0278077 -0.015642 -0.0278077 -0.0251688 0.0678655 -0.0400598 -0.0517723 -0.0335493 -0.00284336 0.0716345 0.0445738 0.0167548 0.0122439 -0.0268469 0.0485904 0.016026 -0.0751815 0.079841 -0.0790154 0.0896523 0.0613015
internal_weight=0 49.7454 39.9908 36.787 9.75459 6.43756 21.7978 15.8466 12.75 3.09662 3.58055 14.9892 5.95116 9.16948 12.6452 10.7709 7.89141 4.54168 3.20376 2.46331 3.31703 3.51142 2.16471 3.97778 2.39719 3.34973 8.123 2.43974 7.00221 2.83583
internal_count=1000 932 755 665 177 122 406 285 233 52 57 259 121 176 224 198 147 75 90 40 55 70 42 68 42 72 155 51 137 68
is_linear=0
shrinkage=0.1
Tree=44
num_leaves=31
num_cat=0
split_feature=0 1 8 7 11 1 10 6 5 0 30 0 28 6 9 4 4 3 5 7 19 4 8 11 29 25 9 2 27 17
split_gain=0.948256 1.41108 1.7293 1.96907 1.27896 1.64801 1.66421 1.459 1.14854 1.15051 1.12808 1.09566 1.42482 1.17243 0.935759 0.872787 0.842445 0.838445 1.37594 0.813644 0.791692 0.923823 1.28275 0.777739 0.673555 0.599645 0.50728 0.40397 0.203572 0.149908
threshold=-1.608339190483093 3.992965936660767 -0.82658091187477101 -1.6642212867736814 4.8875966072082528 -2.1899367570877071 -1.3095270991325376 1.7042843699455263 0.049624711275100715 0.16223448514938357 -1.0374191403388975 0.77826052904129039 -0.68067798018455494 -1.4588729143142698 -1.6665836572647093 -0.6562213897705077 1.4679937958717348 0.37997007369995123 0.25651139020919805 -0.20904585719108579 0.8163239359855653 -1.4721328616142271 1.1326235532760622 3.0093994140625004 -0.092528741806745515 0.5140151083469392 1.444472134113312 -0.18649876117706296 -0.015045809559524058 0.65595531463623058
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=10 2 3 8 5 6 27 11 9 -2 -1 12 15 19 -6 -7 -16 -12 26 -13 21 -14 -23 24 25 -15 -19 -4 -9 -5
right_child=1 -3 4 29 14 7 -8 28 -10 -11 17 13 20 23 16 -17 -18 18 -20 -21 -22 22 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.065335075689466018 -0.013957165900172708 -0.08781098225363182 0.046325345838846177 0.10534600907511008 0.10716942486017432 0.0090361099113185211 -0.027811579639922437 -0.054921555090974787 -0.068695529975343339 0.10188735130897203 -0.098851619410121938 -0.036356519864540984 0.064946586254660127 -0.077188760093404418 0.061697594409106229 -0.11208485178182431 -0.03542682802985813 -0.0056061334375261083 0.0762450985206575 0.06309169040646484 0.088367594407366551 0.028371991960443394 -0.10609216349882994 0.0023632791368135221 -0.10920272376947826 0.018530203681871912 -0.10369224223453927 0.11461046763382945 -0.10733753219287245 0.061704403595753667
leaf_weight=0.94003053195774544 2.0251073632389311 1.5825008619576681 1.6925293132662775 3.366811790969221 1.6007805392146108 1.1306528840214016 2.3436145577579692 1.237226107157767 1.5314348507672546 1.4866825155913828 2.3272018106654278 1.596233665943146 2.100408579222858 1.672293463721872 1.8649891875684255 1.2556327562779186 1.7136793695390222 0.87800183240324337 1.015317842364311 1.6976839303970344 1.9772366024553774 1.9223512243479617 1.1244498295709489 1.6929828003048895 2.8355166940018504 1.0753362774848936 1.3199319252744315 1.7748565040528772 1.8472208287566889 1.0272352397441862
leaf_count=18 36 24 22 147 20 23 35 27 20 34 63 19 35 32 22 22 26 20 12 30 27 34 26 26 47 18 39 20 36 40
internal_value=0 0.00494907 0.00831643 0.0462074 -0.00215291 -0.0103341 0.0372818 -0.0222786 0.00357046 0.0350845 -0.0359551 -0.0124428 0.0121823 -0.0346001 0.0436166 -0.0546962 0.0151886 -0.0531407 -0.0200345 0.0148991 0.0345828 0.0139221 -0.0212532 -0.0570085 -0.0750119 -0.0397274 -0.0645101 0.0812786 -0.0863126 0.0951435
internal_weight=0 45.1754 43.5929 9.43727 34.1557 28.9762 5.811 23.1652 5.04322 3.51179 6.48048 20.0808 9.51073 10.57 5.17945 2.38629 3.57867 5.54045 3.21325 3.29392 7.12445 5.14721 3.0468 7.27613 5.58315 2.74763 2.19793 3.46739 3.08445 4.39405
internal_count=1000 848 824 277 547 479 77 402 90 70 152 339 167 172 68 45 48 134 71 49 122 95 60 123 97 50 59 42 63 187
is_linear=0
shrinkage=0.1
Tree=45
num_leaves=31
num_cat=0
split_feature=8 7 4 11 22 13 2 8 9 19 6 29 11 26 9 4 30 6 10 7 9 1 7 1 4 1 19 17 18 18
split_gain=0.91698 1.4602 1.06126 1.56401 1.36814 0.99755 0.988809 1.24763 0.963603 0.932555 0.919977 1.10619 1.08221 0.872256 0.834856 0.753258 1.04545 1.49972 1.38606 0.798294 0.753758 0.744382 0.709745 0.59234 0.584784 0.566256 0.528305 0.471867 0.814617 0.333562
threshold=0.62742966413497936 2.0546408891677861 0.61790910363197338 0.70849937200546276 0.26117700338363653 -0.45099973678588862 2.5457446575164799 -0.95728960633277882 -1.4426879882812498 0.37843565642833715 0.37987798452377325 -0.4948394894599914 -5.2955746650695792 -0.72909244894981373 -1.6512793302536009 -1.1332715749740598 1.1657342910766604 -0.41582041978836054 -2.5471062660217281 -1.0066578388214109 -0.9361974000930785 0.43324142694473272 -2.0130389928817745 -0.5150548219680785 -1.6112870573997495 2.0035564899444585 0.73767346143722545 0.10223766043782236 -0.088103394955396638 -0.67629358172416676
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 2 3 4 6 9 7 22 -7 21 12 23 -2 -10 -6 20 17 18 19 -17 -14 -4 -1 -12 -9 -5 -13 28 -15 -20
right_child=10 -3 5 25 14 8 -8 24 13 -11 11 26 15 27 -16 16 -18 -19 29 -21 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.015161436007107827 -0.096833810998257469 0.10568960448933658 0.026039555766451644 0.09983780532019812 -0.055494120210887968 0.053524444022043538 -0.1223583569824996 -0.002738811489074012 0.0074928964690683265 0.080684797934374122 0.062521385940556659 -0.095535865465700878 0.0014469945854660334 0.026300260801585082 0.071216114805788336 -0.046276499741698569 -0.095124095502372302 0.07239521453344365 -0.030680186231373197 0.054865487848987329 0.086235638232348283 -0.077588602260593206 0.10209701773480227 -0.039990130146218564 -0.10387056287980563 0.02938984628158009 -0.018210998658578583 -0.10693020264193556 -0.086679705620349323 -0.10325411855750895
leaf_weight=0.99448684975505042 1.1086028628051305 1.5645477757789183 1.8876585243269803 2.8647377025336018 0.63166970014572177 1.030822245404124 0.97122745681554068 0.99408661480993066 1.8923458829522131 1.7058182386681435 1.1211225120350721 3.2978367097675791 2.239832479506731 1.0162438750267027 2.9408485381864011 1.1987928822636607 1.2436607442796219 2.2781049301847824 1.0111735165119151 2.235771747305987 1.9711927846074093 1.0954224383458493 1.0732968216761944 1.133631690405309 1.3458739789202847 1.896196251735091 1.2069578561931846 2.1932194931432605 1.715522894635795 1.6947416299954072
leaf_count=19 22 72 31 76 10 17 16 22 30 26 17 72 36 19 79 20 21 28 25 29 28 20 46 23 29 39 21 36 28 43
internal_value=0 0.0118385 0.00624465 0.0254707 0.0008409 -0.0147839 -0.0310198 -0.0108937 -0.0365872 0.0217098 -0.0155741 -0.0461986 -0.0017569 -0.0502126 0.048812 0.00584063 -0.00954224 0.00310059 -0.0226076 0.0195631 0.0411368 -0.0120139 0.0457023 0.0109813 -0.0609067 0.0717796 -0.0748184 -0.072385 -0.0446501 -0.076134
internal_weight=0 27.814 26.2495 13.7124 8.95149 12.5371 5.37897 4.40774 7.84815 4.6889 21.7414 6.75955 14.9819 6.81733 3.57252 13.8733 9.66225 8.41858 6.14048 3.43456 4.21103 2.98308 2.06778 2.25475 2.33996 4.76093 4.50479 4.92499 2.73177 2.70592
internal_count=1000 615 543 336 221 207 132 116 130 77 385 133 252 113 89 230 166 145 117 49 64 51 65 40 51 115 93 83 47 68
is_linear=0
shrinkage=0.1
Tree=46
num_leaves=31
num_cat=0
split_feature=8 4 11 9 8 6 7 20 1 1 12 22 26 13 12 3 27 9 29 18 4 3 18 30 4 19 4 9 15 3
split_gain=0.86067 1.64068 1.31318 1.72377 1.51715 1.61608 1.3267 1.29167 1.33198 1.1638 1.11832 1.06229 1.02185 0.958074 0.763273 0.719338 0.709592 0.6723 0.632235 0.625436 0.591796 0.55568 0.544829 0.384628 0.338559 0.284887 0.236795 0.513371 0.21752 0.186059
threshold=2.4709751605987553 -3.4333997964859004 -0.97970148921012867 -1.9159966707229612 -1.7129650115966795 2.2315721511840825 2.0273846387863164 0.35997511446475988 0.77400022745132457 -0.80352920293807972 -0.69702842831611622 0.26117700338363653 -0.19718474894762036 -0.59506490826606739 0.58849391341209423 0.62348240613937389 0.52713641524314891 -2.6924784183502193 -0.72692111134529103 -0.2674479186534881 -1.5022985935211179 -0.21885001659393308 -0.36480538547039026 -1.0502278208732603 2.6697212457656865 -0.15976804494857785 -1.0345802903175352 -1.502330422401428 0.17399927228689197 -0.11346042156219481
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 17 6 16 18 7 9 8 10 11 -6 22 -2 -12 21 -15 -4 -1 -5 -14 -10 -9 -3 -11 -20 -23 27 -25 -17 -22
right_child=12 2 3 4 5 -7 -8 14 20 23 13 -13 19 15 -16 28 -18 -19 24 -21 29 25 -24 26 -26 -27 -28 -29 -30 -31
leaf_value=-0.0092841155350607447 -0.10533215179106488 0.013773108830842784 0.094849636412225158 -0.022929158754138226 0.063522395354901229 -0.093482945215413282 0.094731483171740682 0.008098386217068829 0.0091072884722474927 0.010708358735953031 0.038736387678462242 0.056340475028901341 -0.060911895095134075 0.022806249841541216 -0.026962768274742002 -0.049803061300242195 0.010687330243152535 0.097898118228428352 0.1080049856681371 0.022160058848192363 -0.10648145734364448 0.038890967955847476 -0.073428794016335661 -0.10130029315700953 0.03047758820387865 0.10341598368614538 -0.099616563961857987 -0.016963007109022734 -0.1114117038469109 -0.058886076446584369
leaf_weight=0.7645251248031858 1.5828937883488832 1.175657213665545 3.0439999992959192 0.69560890598222647 2.2541746357455903 2.1175465425476423 0.80265267519280392 1.5801245160400865 0.83815762167796481 0.55572195816784997 1.7143988478928855 1.8863054541870949 1.4732525106519463 0.8834437001496539 1.4514451501891015 1.0543105537071826 1.4931934420019386 2.4952282239682964 1.8346038097515731 2.3550875559449187 1.8151042656973007 1.042718196287751 1.8344871420413247 1.3847582563757885 0.81284846737980831 1.9903720230795436 2.4452548986300817 1.5074848704971375 1.2555450852960346 1.5001626815646885
leaf_count=11 43 24 63 17 31 41 40 21 13 14 26 35 27 18 24 27 23 84 59 39 39 23 43 30 12 30 65 38 16 24
internal_value=0 0.00460858 -0.00109214 0.0108528 -0.000330837 -0.0110028 -0.0293026 -0.000953529 -0.0209117 -0.0385296 0.00521622 -0.00249903 -0.037751 -0.0215646 0.036284 -0.0539387 0.0671518 0.0727602 0.0619104 -0.00980836 -0.0659649 0.0561832 -0.0393708 -0.068466 0.0842017 0.0812335 -0.0767093 -0.0573423 -0.0832911 -0.0849445
internal_weight=0 42.2298 38.9701 27.3778 22.8406 19.4975 11.5923 17.38 11.3153 10.7897 7.16187 4.89645 5.41123 4.9077 6.06466 3.1933 4.53719 3.25975 3.34306 3.82834 4.15342 4.61321 3.01014 5.89322 2.64745 3.03309 5.3375 2.89224 2.30986 3.31527
internal_count=1000 891 796 507 421 333 289 292 194 249 118 102 109 87 98 61 86 95 88 66 76 74 67 147 71 53 133 68 43 63
is_linear=0
shrinkage=0.1
Tree=47
num_leaves=31
num_cat=0
split_feature=8 1 6 0 8 7 6 11 9 6 7 6 29 25 31 22 12 17 22 8 9 30 11 4 1 15 0 4 7 16
split_gain=0.777305 1.03681 0.920676 0.893191 1.11027 1.05285 1.02959 0.948169 1.23475 1.21272 1.20461 1.0349 1.01166 0.859616 0.856482 1.08526 1.03211 0.882003 0.747166 0.736679 0.704242 0.669711 0.721539 0.627908 0.591133 0.535505 0.368941 0.276419 0.265808 0.237969
threshold=-3.1696652173995967 3.704098224639893 2.7370777130126958 -1.608339190483093 -0.82658091187477101 -1.6642212867736814 0.18065387010574344 4.8875966072082528 1.2301043272018435 -1.7323389053344724 0.72382399439811718 -1.3626608252525327 -0.4948394894599914 -0.80188676714897145 1.3691098093986513 1.6885804533958437 0.16454584896564486 0.32662111520767217 0.88484010100364696 -0.29845374822616572 -1.6665836572647093 -0.39564204216003412 0.95515820384025585 1.4679937958717348 0.61471280455589306 -0.79492822289466847 -1.7654844522476194 -0.79124709963798512 -4.2261116504669181 -0.10250267386436461
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=26 2 3 13 5 6 -5 8 10 -10 14 -12 -13 -2 15 16 17 21 27 -11 -9 -6 -23 -22 -8 -4 -1 -15 -28 -19
right_child=1 -3 25 4 7 -7 24 20 9 19 11 12 -14 18 -16 -17 -18 29 -20 -21 23 22 -24 -25 -26 -27 28 -29 -30 -31
leaf_value=-0.038126991449431998 0.032065299895869967 -0.081569616680143828 -0.12283930591725048 -0.080249032182740251 -0.056673830058150576 0.090939168049787633 0.098775791064144525 0.11282422804915003 -0.055369513685073216 -0.00024547973941179676 0.086775075837659046 0.058944203370417586 -0.064815079417386776 -0.027553335252463104 0.02823052702196481 0.047743859558786669 -0.089077913962937252 -0.096629127234926754 0.015216012845358874 0.087019368897331598 0.063978175690471414 -0.022250227859162211 0.075346679036346453 -0.025975739488193955 -0.0093516657017370421 -0.023015895179559973 -0.005638544187480366 -0.10494583525607287 0.10099130862913244 -0.034747823664283271
leaf_weight=0.30549991689622413 1.3553858548402784 1.6143002519384015 0.88259086664766107 0.93463741987943683 1.4367222078144539 2.9564829091541425 1.0485708825290205 1.1235191356390712 1.1729025971144436 1.4255973286926744 1.7664347626268861 1.1571445595473051 1.5389837976545087 0.59075859002769027 2.0364065654575816 1.2285334672778834 4.4192121741361916 1.6482611522078516 0.94713055342435826 3.009751101024448 1.6032924298197024 1.5252506854012744 1.5049217175692318 1.5038476977497337 0.97642701771110285 1.3740438888780762 0.28486569877713908 2.1091999686323102 1.3036523805931197 0.99755161907523859
leaf_count=6 27 28 13 16 33 132 24 16 25 21 42 19 30 13 28 13 91 31 20 50 20 35 16 25 20 38 6 77 64 21
internal_value=0 -0.00291574 8.78123e-05 0.00359233 0.00923934 0.048731 0.00656785 0.0012101 -0.00623542 0.0350579 -0.0182598 0.0272803 -0.0116991 -0.0359336 -0.0319942 -0.0416053 -0.051124 -0.0275428 -0.0612043 0.0589709 0.0449746 -0.000441217 0.0262208 0.0204407 0.0466382 -0.0620578 0.0625145 -0.0880121 0.0818696 -0.073298
internal_weight=0 43.8879 42.2736 40.0169 35.0145 5.91612 2.95964 29.0983 24.8677 5.60825 19.2594 4.46256 2.69613 5.00247 14.7969 12.7605 11.5319 7.11271 3.64709 4.43535 4.23066 4.46689 3.03017 3.10714 2.025 2.25663 1.89402 2.69996 1.58852 2.64581
internal_count=1000 924 896 845 708 192 60 516 455 96 359 91 49 137 268 240 227 136 110 71 61 84 51 45 44 51 76 90 70 52
is_linear=0
shrinkage=0.1
Tree=48
num_leaves=31
num_cat=0
split_feature=8 4 6 3 0 30 7 6 6 16 20 20 16 2 10 1 0 0 1 20 7 14 11 19 9 21 28 0 25 9
split_gain=0.789198 1.36048 1.1024 1.09803 1.09432 0.797546 1.23154 1.26218 1.09489 1.08922 0.913254 1.23161 0.826483 0.789985 1.20056 0.973903 0.778973 0.917079 0.721283 0.747014 0.709602 0.631603 0.560838 0.621313 0.534626 0.45346 0.446215 0.431951 0.408984 0.322133
threshold=1.7825397253036501 -3.3526372909545894 0.77008655667305004 -0.24213722348213193 -3.1315989494323726 0.50428083539009105 -2.1902713775634761 1.6017253398895266 -0.59109562635421742 0.053373422473669059 -0.86064061522483815 0.25876171886920935 -1.1816231012344358 2.1508438587188725 -2.5471062660217281 2.2099819183349614 -0.10111981630325316 2.7861731052398686 -1.5318192243576048 0.5318211317062379 -2.9957095384597774 0.11206659302115442 3.2451776266098027 0.13681121915578845 -2.6924784183502193 0.87398377060890209 0.51755946874618541 -0.32855796813964838 0.17315069586038592 2.7368977069854741
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 24 16 13 -5 6 8 12 -6 -9 -7 26 -8 14 18 25 -2 22 -3 -20 -16 -10 23 -18 -1 28 -12 -13 -15 -14
right_child=2 3 -4 4 5 10 7 9 21 -11 11 27 29 15 20 -17 17 -19 19 -21 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.0034424894926538604 -0.080579572968904611 0.07861489843090326 -0.094441966775119873 -0.10492818804502324 -0.088404463951710005 0.077620363798215422 -0.022852777990815133 0.065535843794769386 0.082367983425141136 -0.086523413817285386 -0.031323437044244291 -0.029129262154698434 0.089608803974569606 -0.037092541270779093 -0.0013321573772247966 -0.062969596157456037 -0.043754810335795674 -0.070860325068740723 -0.068512340037039113 0.025187091746079526 -0.096819575535921742 -0.024204926959188414 0.0769513042623447 0.044198090548678209 0.097539719207456577 0.099747298390585312 -0.11192828540867752 0.080569259261684456 0.055709806779811889 0.026712166064731105
leaf_weight=0.69331521727144751 1.2434327914379535 0.89678001403808583 1.8894323767162857 0.72178705502301355 1.3342830669134857 0.92284279875457276 0.85852952953427752 0.83015271509066213 1.0533776748925441 1.0890718754380939 1.3562853978946798 0.57221945468336344 6.0363444630056735 0.81695356871932689 1.0052724722772826 1.1691203024238346 1.5134145431220516 1.219354912638664 1.7211591405794018 1.6826903969049452 3.4462724719196522 1.177967796102166 1.4990816153585909 1.7114482354372738 2.1503296527080229 0.99717071745544661 1.3913050801493225 0.96308494033291914 1.1341543886810539 0.94126722402870644
leaf_count=10 46 14 56 26 24 22 20 23 22 20 31 15 147 22 18 23 28 26 46 25 85 23 14 23 87 15 20 33 22 14
internal_value=0 0.00662306 -0.0264716 0.000753407 0.0158722 0.0205784 0.0335488 0.0519312 -0.0167445 -0.0207508 -0.0126114 -0.0320538 0.0697324 -0.0218591 -0.0388525 0.0142631 -0.00860187 0.00645702 -0.00117152 -0.0221921 -0.0752561 0.0261062 0.0264143 0.00292216 0.072919 0.0448891 -0.0721395 0.0396838 0.0168523 0.0811242
internal_weight=0 34.9617 9.07616 32.1181 19.2485 18.5267 13.321 9.75537 3.56563 1.91922 5.20574 4.28289 7.83614 12.8696 8.75217 4.1174 7.18673 5.9433 4.30063 3.40385 4.45154 2.23135 4.72394 3.22486 2.84364 2.94828 2.74759 1.5353 1.95111 6.97761
internal_count=1000 807 193 710 440 414 293 224 69 43 121 99 181 270 188 82 137 91 85 71 103 45 65 51 97 59 51 48 44 161
is_linear=0
shrinkage=0.1
Tree=49
num_leaves=31
num_cat=0
split_feature=8 3 1 4 9 2 5 11 26 4 7 9 3 8 10 13 25 0 5 18 27 3 27 0 21 22 4 0 30 6
split_gain=0.725642 1.08514 1.00593 1.00433 0.988506 0.848285 1.04613 1.02527 1.05056 0.807989 0.776581 0.771325 1.06924 0.896276 1.20022 0.751948 1.06077 0.715144 0.70961 0.692766 0.636151 0.640272 0.619915 0.655893 0.610738 0.565387 0.479949 0.443548 0.374603 0.221801
threshold=0.84256586432456981 -2.818352103233337 1.2805632352828982 -0.29410147666931147 -1.4426879882812498 3.5605782270431523 -0.93297564983367909 -3.1654969453811641 -0.39900171756744379 0.92876344919204723 -1.6642212867736814 -0.141118049621582 1.4703543782234194 -0.82658091187477101 -1.5789780020713804 -1.1710312962532041 -0.94113266468048085 0.30448183417320257 2.5333309173583989 0.11730988323688508 -0.56510543823242176 0.74455237388610851 0.34618109464645391 0.75287938117980968 0.51193365454673778 0.47718013823032385 1.5504462718963625 -0.43635436892509455 -0.4216609895229339 2.0313966274261479
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 -1 5 11 24 6 9 25 -9 10 -2 12 13 27 -15 -6 26 -4 29 -12 21 -18 -22 -24 -5 -8 -17 -3 -10 -13
right_child=2 3 17 4 15 -7 7 8 28 -11 19 18 -14 14 -16 16 20 -19 -20 -21 22 -23 23 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.077016802687213601 0.050634499252129019 -0.0060986385753942184 0.0041772463274558479 0.10858719618230504 0.040954250486055221 0.081207291089228473 -0.084480537406297698 -0.036863561432779017 0.10486914036183551 -0.10767074542470673 -0.10550133681119792 0.097675489368968804 0.10441903898355441 0.032254574460641665 -0.092453629520270614 0.080985193787074597 -0.042938252015690098 -0.10156349317419618 -0.036013013613807869 -0.0016332100173711132 -0.090443699537717484 0.060435598202046796 -0.062758504632254811 0.043864217965773444 0.00066952422890790794 0.017899934238079191 -0.011530974675621315 0.10452430932447183 0.03935933023375468 0.029327181388769996
leaf_weight=1.3490189518779541 1.0689194425940516 0.62314471695572138 0.9247465524822468 1.194884135387839 2.1655949009582391 1.0540091190487135 1.3576892223209145 1.3802631776779888 1.4566944846883414 1.5269831321202216 1.2182723060250285 3.1503701745532466 1.0940637136809526 1.4087517033331118 1.7067127553746086 1.1716696647927149 1.4015481481328591 2.0742629645392299 0.53307955618947733 1.357801050879061 3.0175918079912645 1.0465653203427789 1.5407591676339509 0.922304531559348 0.93457154324278224 0.89496408961713303 1.0754034705460069 0.86638470692560277 2.1779962275177263 0.55905149690806855
leaf_count=38 14 10 20 30 37 14 34 29 16 45 37 115 62 26 36 21 30 50 15 32 56 19 36 22 17 16 19 43 38 23
internal_value=0 0.0102914 -0.0165723 0.015116 -0.00169577 -0.00492931 -0.0122277 0.0122361 0.0374092 -0.0466039 -0.0210214 0.0395871 0.0155545 -0.00555813 -0.036063 -0.0125524 -0.0239395 -0.0689583 0.0718707 -0.0507544 -0.0411277 0.00125393 -0.0600588 -0.0228332 0.0612245 -0.0438055 0.0367088 0.0582452 0.065614 0.0873746
internal_weight=0 25.7615 16.4926 24.4125 14.4709 13.4936 12.4396 7.26761 5.01495 5.17198 3.64499 9.94156 5.69906 4.60499 3.11546 12.3414 10.1758 2.99901 4.2425 2.57607 7.92877 2.44811 5.48066 2.46306 2.12946 2.25265 2.24707 1.48953 3.63469 3.70942
internal_count=1000 655 345 617 287 275 261 133 83 128 83 330 177 115 62 240 203 70 153 69 163 49 114 58 47 50 40 53 54 138
is_linear=0
shrinkage=0.1
Tree=50
num_leaves=31
num_cat=0
split_feature=0 1 8 7 1 10 11 6 15 22 31 24 12 30 22 15 6 4 9 3 5 30 17 1 17 7 11 8 15 18
split_gain=0.670394 0.857874 1.04346 1.33477 0.907858 1.40079 1.07167 1.04848 0.83048 0.734152 1.14689 0.975793 1.12189 0.727343 0.679872 0.665667 0.625127 0.57745 0.572431 0.571085 0.869264 0.557432 0.455725 0.450755 0.382554 0.337049 0.230746 0.210104 0.179874 0.15092
threshold=-1.608339190483093 3.992965936660767 -0.82658091187477101 -1.6642212867736814 -2.3026347160339351 -1.3095270991325376 4.8875966072082528 1.7042843699455263 -0.82525604963302601 1.6885804533958437 1.3691098093986513 -0.62654510140418995 0.36355853080749517 -1.0374191403388975 -0.91528546810150135 0.14212975651025775 0.37987798452377325 -0.44345408678054804 0.24192273616790774 0.37997007369995123 0.25651139020919805 0.66232308745384227 1.0000000180025095e-35 0.82854908704757702 -0.46242572367191309 0.39549702405929571 -1.0592629313468931 0.94824668765068065 -0.94553837180137623 0.40455217659473425
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=13 2 3 8 5 26 7 9 -2 10 11 27 14 -1 -13 17 -10 22 -8 -15 23 -18 -16 -21 -14 -17 -4 -6 -5 -9
right_child=1 -3 4 28 6 -7 18 29 16 -11 -12 12 24 19 15 25 21 -19 -20 20 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.055260340370468924 -0.068010297439199827 -0.077480480042209907 0.053457473607528644 0.048530569913263009 -0.052389982472606789 -0.032645899724691015 0.075085232551867997 -0.10130011637950841 -0.045354427099045622 0.05653886799920959 0.053127879213758146 -0.064729245407690192 -0.0040921555597411311 -0.095866174038997248 0.070018330488662892 0.029881566193844619 0.098705283955177678 -0.070407848781550403 -0.0016594740457169022 -0.099380704416163715 0.065573622863117895 -0.0084620564108520667 -0.034020309427989164 0.0069948744440139158 -0.083508155153374175 0.099135504117629392 0.11047560055840872 -0.1055781828381645 0.10425863501691775 -0.048987954360504543
leaf_weight=0.76265976345166553 1.1577521748840807 1.2289870437234629 1.3995582750067115 0.76780442567542229 1.5194224035367416 1.6884587462991465 2.0211209133267398 1.4799836943857365 0.76167644467204842 1.485331967473029 2.0077443104237309 1.0397748490795535 0.82288650516420636 1.6891028177924443 0.9452590839937367 1.8868762757629158 1.3552249763160946 1.0567752821370984 1.872213485650718 1.0955370836891236 0.79778305441141117 0.75618442008271802 0.75917790411040187 0.62593605555593956 2.3073522513732314 1.1198264663107691 1.4400328900665045 1.4528014874085782 2.3577095477376089 0.87906191591173399
leaf_count=18 19 24 22 34 36 31 31 42 15 17 33 28 18 63 19 37 38 21 34 48 12 18 21 11 53 21 21 40 153 22
internal_value=0 0.00461185 0.00754983 0.0415237 -0.0013941 0.0394837 -0.00956411 -0.0194715 0.00349419 -0.0105067 -0.0171823 -0.0281167 -0.0130816 -0.0345938 0.00970171 0.0231193 0.032308 -0.0123306 0.0381805 -0.0508776 -0.0207138 0.0603242 0.0236782 -0.060702 -0.062631 0.0556747 0.0823729 -0.078388 0.0905686 -0.0818068
internal_weight=0 35.569 34.34 7.15635 27.1837 4.52805 22.6556 18.7623 4.03084 16.4032 14.9179 12.9102 9.93793 4.97102 6.80769 5.76792 2.87309 2.76121 3.89333 4.20836 2.51926 2.11141 1.70444 1.72147 3.13024 3.0067 2.83959 2.97222 3.12551 2.35905
internal_count=1000 848 824 277 547 74 473 408 90 344 327 294 218 152 147 119 71 61 65 134 71 56 40 59 71 58 43 76 187 64
is_linear=0
shrinkage=0.1
Tree=51
num_leaves=31
num_cat=0
split_feature=8 4 0 3 1 11 26 9 2 30 20 26 6 7 5 6 20 3 16 2 28 9 10 14 1 4 25 25 0 7
split_gain=0.663026 1.10835 0.894174 0.845537 1.06021 1.71074 0.988588 0.874755 0.872678 0.83345 1.0119 0.807727 0.788404 0.981997 0.860095 0.811928 0.738736 0.69923 0.63832 0.613627 0.537874 0.521544 0.484837 0.400342 0.406347 0.394239 0.35283 0.825362 0.617706 0.240491
threshold=2.4709751605987553 -3.4333997964859004 -3.1315989494323726 -0.22925552725791928 0.46325501799583441 0.40085124969482427 -0.051145128905773156 -1.3711115121841428 4.1532464027404794 0.50428083539009105 -1.0153217911720274 -0.19718474894762036 1.7042843699455263 -2.8339403867721553 -2.7725118398666377 -0.78662157058715809 0.63155561685562145 -1.2833063602447508 -1.0000000180025095e-35 2.17708432674408 0.24971261620521548 -2.6924784183502193 -2.8759547472000118 0.064470633864402785 -0.87797400355339039 -0.1948781609535217 0.10375598073005678 -0.62472391128540028 -0.37390863895416254 -1.7502976059913633
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 21 -3 4 5 17 7 -6 9 12 -11 -2 13 15 -15 -5 25 29 -14 -9 -18 -1 -19 -13 -25 -12 27 -16 -29 -4
right_child=11 2 3 8 6 -7 -8 19 -10 10 16 23 18 14 26 -17 20 22 -20 -21 -22 -23 -24 24 -26 -27 -28 28 -30 -31
leaf_value=-0.0096128203896763027 -0.10433715364763373 -0.10367609018597489 -0.1063731029019036 -0.094388327289884333 0.057335415315284161 0.07784099359545292 -0.090189868150340496 -0.095866449231284068 0.10838282417013781 0.088912599462114911 -0.015031202905430516 -0.059048158565855338 0.04091776746086441 -0.047576665849202393 0.080308603824629857 0.029513765939262627 -0.040036740972672381 0.059882383402339694 -0.083918970492524336 0.0057520106887559005 0.089447578517966364 0.093579763751015566 -0.043582165306580808 0.06221114933894592 -0.025866016865529071 -0.09301204378246325 0.084236694897439818 -0.071156689102311071 0.041040689082944337 -0.035860786446437157
leaf_weight=0.65602983348071608 1.2592984689399598 0.82008965639397757 0.99543435219675303 0.96082528028637204 1.2236861642450096 2.397632178850472 2.8501846226863563 1.228355926461518 0.96344324853271146 0.68611968331970263 0.94029081822372962 1.0858133267611259 0.65019255271181475 0.82541982876137332 2.4510362353175914 1.1764658577740235 0.67321736970916368 1.022070496343076 1.1068861479870962 1.151097424328327 0.61285058874636889 1.9325727995019395 0.81331668514758515 0.99257602216675911 1.1091072172857819 2.0878304783254857 3.3587875089142472 0.8461577678099278 1.1681010099127807 0.94087058119475864
leaf_count=11 43 41 26 19 18 35 77 35 22 19 35 26 22 23 55 23 23 20 24 26 18 84 28 16 24 39 92 18 34 24
internal_value=0 0.00449322 -0.000605699 0.00210957 -0.0178465 0.0117945 -0.0461831 -0.0113724 0.0157201 0.0106315 -0.0238901 -0.036531 0.0243927 0.034511 0.0495094 -0.0261868 -0.04183 -0.0301907 -0.0377242 -0.0467069 0.0216665 0.0674277 0.014034 -0.00974249 0.0157308 -0.0687975 0.0597517 0.0413342 -0.00609163 -0.0721104
internal_weight=0 34.539 31.9504 31.1303 12.6226 6.16932 6.45332 3.60314 18.5076 17.5442 5.00031 4.4468 12.5439 10.7868 8.6495 2.13729 4.31419 3.77169 1.75708 2.37945 1.28607 2.5886 1.83539 3.1875 2.10168 3.02812 7.82408 4.4653 2.01426 1.9363
internal_count=1000 891 796 755 289 133 156 79 466 444 134 109 310 264 222 42 115 98 46 61 41 95 48 66 40 74 199 107 52 50
is_linear=0
shrinkage=0.1
Tree=52
num_leaves=31
num_cat=0
split_feature=20 8 24 4 9 11 9 1 1 9 4 5 7 3 2 19 5 15 16 9 22 11 10 15 21 31 29 7 9 13
split_gain=0.618181 0.703005 1.01517 0.995634 1.02188 0.959586 0.954076 0.903523 0.884776 0.881147 0.978875 0.833868 0.976647 0.714087 0.667169 0.63237 0.617264 0.73043 0.572305 0.570217 0.60954 0.554476 0.534831 0.444849 0.440312 0.419905 0.320443 0.238798 0.177536 0.158235
threshold=1.9680437445640566 1.1326235532760622 -0.46206876635551447 0.4314530491828919 -1.4426879882812498 0.70849937200546276 0.26169002056121832 1.2805632352828982 0.47342592477798467 -2.2834039926528926 -3.3526372909545894 -0.51409405469894398 -1.3801147341728208 -0.40226960182189936 2.0050599575042729 0.076233550906181349 -0.93297564983367909 0.14212975651025775 0.2598749846220017 2.3834861516952519 0.257268637418747 -0.38942228257656092 2.4017443656921391 -0.35176825523376459 0.071496769785881056 -0.023028294555842873 -0.12295129895210265 -0.66305851936340321 -0.47277516126632685 -0.14496307820081708
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 3 14 5 -5 9 -6 16 11 -1 -11 12 -8 -7 -3 -10 17 27 26 20 23 -13 29 24 -12 -25 -18 -4 -9 -22
right_child=-2 2 7 4 6 13 8 28 15 10 19 21 -14 -15 -16 -17 18 -19 -20 -21 22 -23 -24 25 -26 -27 -28 -29 -30 -31
leaf_value=-0.074294544463181525 0.074176422294265798 -0.098661996904963037 -0.026727586834252238 0.066927585557666258 -0.083511931680878101 0.015548045501962753 0.036569804170246321 -0.10410106694478523 -0.10750770369881668 0.095767996282363321 0.081624375181525016 0.0090105457889680125 -0.082462502011606043 0.097718504801926995 -0.00062725345205815169 -0.011347544132350847 0.09678199663914705 0.036513552090293361 -0.0238179068102254 0.075562382896751989 0.029880551877327508 0.098757147587583444 -0.038828676903534143 -0.01622512411383811 -0.038452811500209988 -0.10820047984902664 0.029984712110169166 -0.10456636249158908 -0.034679803730722013 0.10715835483260823
leaf_weight=1.1587833748199008 1.084982216823845 2.3307123910635705 0.72711062105372437 1.3843941781669862 2.1705967513844344 1.7800477375276385 1.3380997586064021 0.8532257080078125 1.2592940023168924 1.3693522471003188 0.5092258616350549 1.4036634089425208 1.4216272318735721 2.6058073095045975 0.98864748049527396 1.4966811640188096 1.3024627240374684 1.1210090448148546 1.1342001645825801 0.89203748619183887 0.42848852509632696 1.3509810157120226 0.6072130724787711 0.91666961670852742 0.76286189910024405 1.0825922419317251 1.6009396910667411 0.86063695978373334 0.64828185550868489 0.69431626936420787
leaf_count=25 18 75 19 35 47 33 31 27 30 77 19 31 27 99 24 32 18 25 27 25 22 26 21 33 27 28 29 32 15 23
internal_value=0 -0.00242895 -0.022765 0.00712061 -0.0138021 0.026439 -0.0245062 -0.00397135 -0.00901988 0.00668551 0.019606 0.0141031 -0.0247477 0.0643687 -0.0694631 -0.0552862 0.0116431 -0.0252867 0.0364187 0.0019095 -0.0112271 0.0530256 0.0367788 -0.0366146 0.00961495 -0.0660294 0.0599499 -0.06892 -0.0741282 0.0776673
internal_weight=0 36.2 11.5672 24.6327 11.8253 12.8074 10.4409 8.24787 8.27035 8.42154 7.26276 5.51437 2.75973 4.38586 3.31936 2.75598 6.74636 2.70876 4.0376 5.8934 5.00137 2.75464 1.73002 3.27135 1.27209 1.99926 2.9034 1.58775 1.50151 1.1228
internal_count=1000 982 291 691 259 432 224 192 177 300 275 115 58 132 99 62 150 76 74 198 173 57 66 107 46 61 47 51 42 45
is_linear=0
shrinkage=0.1
Tree=53
num_leaves=31
num_cat=0
split_feature=6 0 15 3 8 2 5 1 11 4 9 9 8 13 2 13 26 11 8 16 10 4 7 21 31 19 1 7 15 6
split_gain=0.625539 0.862834 1.32261 0.690933 0.897318 0.655485 0.765138 0.736963 1.01045 0.652555 0.804693 0.652756 0.695671 0.69118 0.622959 0.588186 0.779812 0.526219 0.508962 0.496282 0.616775 0.496371 0.492885 0.448281 0.42572 0.318081 0.308359 0.298687 0.257445 0.334428
threshold=2.2315721511840825 0.46007883548736578 -0.8999110460281371 -2.7369303703308101 0.93145474791526806 3.0034747123718266 -0.93297564983367909 0.43324142694473272 -3.1654969453811641 -0.16142201423645017 -1.3997637033462522 -0.4592334628105163 -0.79147225618362416 1.3032695651054385 -0.10519874095916747 -0.44223669171333307 -0.35152202844619745 -0.86672562360763539 0.016810417175292972 0.54110366106033336 1.739203989505768 -0.97413113713264454 -1.5177835822105405 -0.060760401189327233 0.78099998831748974 -0.46053160727024073 0.43324142694473272 0.26619952917099005 0.42414110898971563 0.32099404931068426
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=3 -2 -3 -1 9 6 19 8 25 11 -11 12 -5 -13 -14 17 23 -12 -4 20 -6 -22 -16 -17 -10 -8 -19 -9 29 -18
right_child=1 2 18 4 5 -7 7 27 24 10 15 13 14 -15 22 16 28 26 -20 -21 21 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.062716474783969176 -0.096781331392035527 -0.11158116704650148 0.099354562435815086 0.077768454802135858 -0.10524474261676633 0.054018572327237717 0.0043949475042689267 -0.104333514953954 0.022524178107276662 0.089217158932685239 -0.03514238144936338 0.094345759523456266 -0.078878805569914903 -0.013724494105594493 -0.042432698325925472 -0.023909847688788663 0.0017174632852754305 0.078148466473387149 -0.022850427918475587 -0.0016795694045917044 0.065539487435467975 -0.07078641609016198 0.058233232974910745 0.061617738746117139 0.09557035316019083 -0.083535348280464702 0.0076544497282128926 -0.015154571397552047 -0.094680001817255055 -0.083493204786819014
leaf_weight=1.4680818943306793 1.5256117065437131 0.70312906522303853 0.94071304798126198 1.2821452387142924 2.1244188412092622 1.2979386160150159 0.74050889117643237 0.76541818957775809 2.0632976246997723 1.0160143561661232 0.99029172724112902 3.0343619699124269 0.91749935969710383 0.73519327770918597 0.78261770028621014 1.2824204359203579 1.0570659646764435 1.5823085275478654 0.53441820200532675 1.3541430155746637 0.4648451954126358 0.62779807392507769 1.2849869071505959 1.1737006190232921 1.3009222298860548 0.92564943199977123 1.0208499468863008 0.7373872771859169 1.2038357788696883 0.81625183299183834
leaf_count=41 55 11 46 70 64 22 17 25 38 32 28 141 18 18 18 25 30 34 14 33 8 32 31 27 19 26 16 16 26 19
internal_value=0 -0.0391089 0.00128382 0.00429683 0.00751381 -0.0132249 -0.0210846 0.000872495 0.0192299 0.021662 0.00479721 0.042946 0.0161602 0.0732683 -0.0103014 -0.00460067 -0.0250587 0.0269011 0.0550815 -0.0524658 -0.0738431 -0.012789 0.0201297 0.016961 0.0507707 -0.0444555 0.0505037 -0.0605757 -0.058598 -0.035411
internal_weight=0 3.70387 2.17826 32.05 30.5819 12.4023 11.1044 6.53318 5.03038 18.1795 10.1427 8.0368 4.26725 3.76956 2.9851 9.12672 5.53327 3.59345 1.47513 4.57121 3.21706 1.09264 2.0676 2.45612 3.36422 1.66616 2.60316 1.50281 3.07715 1.87332
internal_count=1000 126 71 874 833 300 278 141 100 533 237 296 137 159 67 205 127 78 60 137 104 40 49 52 57 43 50 41 75 49
is_linear=0
shrinkage=0.1
Tree=54
num_leaves=31
num_cat=0
split_feature=6 0 2 8 10 6 7 11 7 9 4 6 9 5 1 8 1 26 6 3 31 25 12 17 11 26 15 9 7 25
split_gain=0.558005 1.01111 1.11446 0.613845 0.801982 0.78503 0.859197 0.720449 0.749547 1.06833 1.34616 1.11434 0.722695 0.776356 0.620565 0.593354 0.578831 0.548386 0.557069 0.545041 0.469544 0.864035 0.647913 0.603735 0.369337 0.300931 0.282372 0.250337 0.190445 0.179534
threshold=2.090929508209229 0.44619986414909368 1.8257751464843752 -1.1515020728111265 4.2750437259674081 -4.106367588043212 -2.1114803552627559 3.7105647325515752 0.88489025831222545 0.3670168817043305 -1.6112870573997495 -2.0841283798217769 1.1047214865684511 -0.57589870691299427 2.1486637592315678 2.9477462768554692 0.4226242601871491 -0.093423571437597261 -0.80244350433349598 1.1476703882217409 0.92558419704437267 -0.062490180134773247 0.079420387744903578 -0.67083266377449025 1.8407537937164309 -0.32592111825942988 -0.8999110460281371 2.8944664001464848 0.25074130296707159 -0.22630245238542554
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=3 -2 26 5 7 -1 16 8 9 20 -11 -12 25 -14 19 17 -7 -10 -19 -9 22 -22 23 -5 -25 -13 -3 -8 -24 -4
right_child=1 2 29 4 -6 6 27 14 15 10 11 12 13 -15 -16 -17 -18 18 -20 -21 21 -23 28 24 -26 -27 -28 -29 -30 -31
leaf_value=-0.13348618010052055 -0.094452608815489794 -0.1155772557452405 0.030059783851877259 0.029185899869744389 -0.08592858059388514 0.053966713175259889 0.10283057319187588 0.085134845508108681 0.10158537114374762 0.095061355449361729 -0.099585410016950246 0.0077149155875814827 -0.02010751679601978 0.084271079246133238 -0.025360174236276158 -0.045743566633363944 -0.046985322417074607 0.080818329860229507 -0.046138966973739003 0.0042486813618589914 0.067769145668258396 -0.062253480572573884 -0.099825924999369625 -0.09335233697522316 -0.0035969342325324208 -0.07628315859457975 -0.018987287818411608 0.0094923086883302206 -0.042595034003117642 0.1033197636116357
leaf_weight=0.25542836170643557 1.6798669663257886 0.56554200127720833 0.64085843181237556 1.3448050096631052 1.0815575832966704 1.1226327517069878 1.9900579381501331 2.3889920641668136 1.3642031128983951 1.54305651364848 1.1209689332172272 0.75432490278035402 1.2061954303644593 1.741299594752491 1.2216547154821453 0.71424156101420511 1.1495465170592067 0.64090336253866587 0.75013876054435824 1.2791092367842791 1.0048900386318567 1.040057382546365 2.3733277088031177 1.1690966608002868 0.75423201732337464 0.98142391722649225 0.65112459845840953 0.33583819121122349 0.77011856576427806 0.69978476362302899
leaf_count=4 65 8 19 24 43 32 133 37 37 36 48 17 33 32 36 20 33 24 22 27 14 27 66 37 15 24 34 9 20 24
internal_value=0 -0.0341806 0.00541146 0.00457925 -0.00168261 0.0371491 0.0466281 0.0020883 -0.00660914 -0.0158492 0.0120445 -0.0100257 0.0114111 0.0415565 0.0363698 0.03548 0.00289281 0.0565355 0.0123548 0.0569289 -0.0400839 0.00163982 -0.0533916 -0.022215 -0.0581548 -0.0397791 -0.0638851 0.0893534 -0.0858048 0.0682998
internal_weight=0 4.23718 2.55731 30.0981 25.2446 4.8535 4.59808 24.163 19.2733 15.8038 7.34727 5.80421 4.68324 2.9475 4.88976 3.46949 2.27218 2.75525 1.39104 3.6681 8.45653 2.04495 6.41158 3.26813 1.92333 1.73575 1.21667 2.3259 3.14345 1.34064
internal_count=1000 150 85 850 639 211 207 596 496 393 190 154 106 65 100 103 65 83 46 64 203 41 162 76 52 41 42 142 86 43
is_linear=0
shrinkage=0.1
Tree=55
num_leaves=31
num_cat=0
split_feature=0 9 8 1 28 23 8 7 2 4 3 6 1 15 2 22 25 1 4 24 12 24 7 1 25 26 5 25 0 28
split_gain=0.542256 0.637858 0.579763 0.765243 0.58728 0.731742 0.570822 0.708198 0.745461 0.620108 0.628687 0.594174 0.806655 0.537161 0.535412 0.572326 0.582489 0.492381 0.460223 0.426168 0.863969 0.707906 0.341232 0.519891 0.568207 0.271568 0.274047 0.201229 0.368791 0.154674
threshold=-1.608339190483093 3.3240535259246831 -3.0854187011718746 3.704098224639893 -1.1529585123062132 -0.77515411376953114 0.62742966413497936 -0.86519092321395863 -1.6019626855850218 0.90728291869163524 -1.429771304130554 0.40676501393318182 -0.5150548219680785 -0.97393241524696339 1.4031839370727541 0.88484010100364696 -0.80188676714897145 0.082951605319976821 -0.86077696084976185 -0.62654510140418995 0.84090170264244091 -1.2798794507980344 0.72382399439811718 0.31563207507133489 -0.0068046730011701575 0.13146134465932849 -0.95051786303520192 0.48559746146202093 0.50370800495147716 -0.20270609110593793
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=14 2 -2 4 5 -4 7 9 -9 10 -6 19 -13 -10 15 16 -1 -11 -18 21 22 -8 23 24 -21 -12 -27 28 -7 -14
right_child=1 -3 3 -5 6 27 11 8 13 17 25 12 29 -15 -16 -17 18 -19 -20 20 -22 -23 -24 -25 -26 26 -28 -29 -30 -31
leaf_value=0.039680439253599278 0.077153544439701904 0.083996041175399855 0.034394185775857657 -0.085286874973678567 -0.027763855703107007 -0.092429618021660109 0.028901542046458423 -0.054138974581551126 -0.0099264882925563635 0.014872787436294633 0.094437675972450341 0.0083710531106818663 -0.044393775312482724 0.091299529677621893 -0.091316148182987097 0.07026273296468416 0.028142283303969836 -0.071737868930574253 -0.10358997715254509 0.086811452066112571 -0.05578749125823345 -0.084619107813677416 0.084388593899632838 -0.034898821819196263 -0.0063849221247674727 0.076965603224519924 -0.017109729092696455 -0.10510529031415086 0.0087788282749451177 -0.10507498620006384
leaf_weight=0.68820817954838354 0.98493730055634188 0.97716840077191491 1.0092238970100882 1.0347549975849677 1.2157890982925894 0.76094471290708587 1.0642809569835665 0.52334850421176282 0.63578620599583757 1.1973098437301817 1.1901956237852673 1.7803996331058445 0.62440047040581714 2.98766041058113 1.1791307558305559 0.66201490303501476 0.34891903493553478 1.4528777892701326 1.1053916844539329 1.5630288403481236 1.1710839224979279 1.1352952383458603 1.2704958431422708 1.0597403878346083 1.1251028669066732 0.58996100258082174 0.65171573613770262 0.88472048146650184 0.68336986633948982 1.283554877620189
leaf_count=17 69 24 27 25 32 22 21 20 31 24 36 38 22 138 56 16 11 31 52 22 25 34 22 24 29 22 21 23 26 40
internal_value=0 0.00458211 0.00179873 -0.000960882 0.00241328 -0.0367293 0.00821507 0.0253344 0.0574244 0.00420482 0.0309514 -0.00658988 -0.0400409 0.073538 -0.0347766 -0.0110053 -0.0361162 -0.0326086 -0.0719847 0.00811734 0.0215536 -0.0296913 0.0396019 0.0244196 0.0478046 0.0603055 0.0275885 -0.0675488 -0.0445434 -0.0852164
internal_weight=0 28.8571 27.88 26.895 25.8603 3.33826 22.522 10.4446 4.1468 6.29785 3.64766 12.0774 3.68835 3.62345 3.98366 2.80453 2.14252 2.65019 1.45431 8.38903 6.18945 2.19958 5.01837 3.74787 2.68813 2.43187 1.24168 2.32904 1.44431 1.90796
internal_count=1000 848 824 755 730 98 632 355 189 166 111 277 100 169 152 96 80 55 63 177 122 55 97 75 51 79 43 71 48 62
is_linear=0
shrinkage=0.1
Tree=56
num_leaves=31
num_cat=0
split_feature=13 11 3 8 5 7 9 4 0 5 4 8 11 14 29 11 5 25 10 10 7 29 4 1 5 29 22 13 29 10
split_gain=0.535235 0.692063 0.642426 0.56326 0.7138 0.708242 0.709376 0.687328 0.709079 0.664042 0.706783 0.68077 0.747781 0.662409 0.512725 0.49669 0.668844 0.521764 0.488531 0.828565 0.629384 0.470305 0.666026 0.468281 0.460437 0.410265 0.398007 0.397905 0.362522 0.280548
threshold=1.4291877746582033 3.3369733095169072 -2.818352103233337 0.93145474791526806 -2.1957887411117549 -1.6642212867736814 0.68402928113937389 0.45788866281509405 0.23862367868423465 0.37042063474655157 -1.8932023048400877 -0.75641953945159901 4.6643924713134775 1.3423827886581423 -0.092528741806745515 -5.2955746650695792 -0.87853625416755665 -0.47056014835834498 -4.9047172069549552 -2.084357380867004 -1.8983382582664488 -0.1444931477308273 0.74304640293121349 0.81392666697502147 -0.29206293821334833 0.54209741950035106 0.057140542194247253 1.5627481341362002 -1.0039590597152708 0.34161166846752172
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=2 23 -1 5 28 6 24 -8 29 11 26 21 14 15 -13 -6 17 -17 -18 -20 -21 22 -7 27 -4 -16 -11 -2 -5 -9
right_child=1 -3 3 4 13 9 7 8 -10 10 -12 12 -14 -15 25 16 18 -19 19 20 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.076434374149310355 0.09701834911074321 0.11003227542909325 -0.010289837881272993 0.011907171958072539 -0.10544841364732332 0.086947532927025756 0.081590948057072482 -0.0054787456708537026 0.046773877291349256 0.06569270697471595 0.094928625632382441 0.030931795180407919 0.073037946968405446 0.068969533558114415 -0.092184194317810469 -0.0017258119272524773 -0.048105478353309333 -0.0999674963224743 0.076079500629007354 0.059844198596900905 -0.06648928172297934 0.098358957890506915 -0.078112888478456224 -0.042776605879542927 -0.08257042385821893 -0.0080757400542494032 -0.071502293640038739 -0.0029277047058310357 -0.10300982673783975 -0.090676048267115061
leaf_weight=1.1995381410233736 0.74779008375480749 0.93811690900474776 1.7867033658549192 0.34479150781407941 0.60463502677157421 0.48618516908028042 1.1319012627936902 0.66775753023102902 1.2135351132601497 0.37535205716267261 2.2615108476020396 0.85148399882018599 0.7804110487923025 0.97814062144607294 1.4799879798665512 0.97239075228571925 1.0775199923664316 1.2174690733663718 1.8069698140025123 0.61616368312388514 1.0954193156212619 1.1462845921050757 0.49168000370264053 1.025716657284647 1.7391610285267329 0.95362872723489989 0.48425791668705642 0.85238747298717499 1.3468173160217691 0.91765454271808267
leaf_count=43 16 15 45 13 21 43 46 18 25 29 97 33 20 20 32 19 24 39 30 12 44 64 11 29 45 25 19 28 70 25
internal_value=0 0.0363075 -0.00483553 -0.00163416 -0.0203406 0.00958947 -0.013376 0.0158355 -0.0107562 0.0279818 0.06559 0.00901797 -0.0149539 -0.00836479 -0.0358573 -0.0186 -0.0108617 -0.056344 0.010809 0.0288509 -0.0210097 0.0548989 0.00395357 0.00996885 -0.0459428 -0.0592257 -0.0115956 0.0437788 -0.0795869 -0.0547919
internal_weight=0 3.56401 28.0274 26.8278 10.0603 16.7675 7.45671 3.93085 2.79895 9.31078 3.12112 6.18966 4.06551 8.36871 3.2851 7.39057 6.78593 2.18986 4.59607 3.51855 1.71158 2.12415 0.977865 2.62589 3.52586 2.43362 0.85961 1.60018 1.69161 1.58541
internal_count=1000 88 912 869 292 577 204 114 68 373 145 228 110 209 90 189 168 58 110 86 56 118 54 73 90 57 48 44 83 43
is_linear=0
shrinkage=0.1
Tree=57
num_leaves=31
num_cat=0
split_feature=20 5 3 5 11 4 9 15 30 20 1 26 23 9 6 6 28 20 8 4 7 14 15 16 4 8 2 30 0 7
split_gain=0.460947 0.500325 0.682581 0.808217 0.763431 0.773285 1.18741 1.17618 0.60651 0.730467 0.677581 0.52279 0.476925 0.458088 0.82339 0.734415 0.642432 0.39518 0.330328 0.336891 0.303394 0.386173 0.281603 0.344551 0.343072 0.173691 0.0902378 0.0780336 0.0328713 0.0224943
threshold=1.9680437445640566 -2.8203281164169307 2.6622943878173833 -0.67728805541992176 2.8430706262588505 -1.0175390839576719 -1.5438122153282163 1.3031024336814883 -0.65799501538276661 0.11574089527130128 0.86920762062072765 -0.82819029688835133 -0.29447802901268 1.2716721296310427 -1.2840110063552854 -1.9400438070297239 -0.014420712832361458 -0.60778766870498646 0.063155800104141249 -2.1543085575103755 -0.17649734020233152 -0.42796327173709864 -1.1510164141654966 0.42295156419277197 0.4314530491828919 0.14835035800933841 2.2022285461425786 -1.3599413037300108 -0.83598113059997547 0.25074130296707159
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 17 4 12 5 6 7 20 -6 10 -10 -8 -4 15 -15 -7 -16 -1 27 -20 -3 -22 -17 26 -25 -19 29 -13 -5 -24
right_child=-2 2 3 28 8 13 11 -9 9 -11 -12 18 -14 14 16 22 -18 25 19 -21 21 -23 23 24 -26 -27 -28 -29 -30 -31
leaf_value=0.0089800090966617903 0.069608715307239374 -0.099502644184030833 0.044101386505517495 0.072392622305770621 -0.027413306441312124 0.025413671988573985 -0.02581565788815425 0.090655544943233288 0.051461803234834652 0.094844315970676843 -0.050724686671544596 0.034651719210603268 -0.072871402949634942 -0.091601583381146601 -0.035268112764028019 -0.014320129299657675 0.073348900261831759 -0.037712312833388634 0.068593940460732902 -0.012017699507341649 0.04924644227347913 -0.07854369348713526 -0.10564237633243967 0.023345642501715946 -0.073288388645031566 -0.099127925065557607 -0.059339396930759153 0.10362171888095631 0.10521782299451042 -0.085098330953104137
leaf_weight=0.97851989138871465 0.91749177966266771 1.4168674666434531 0.78486365312710404 0.41529175080358971 1.5480943357106296 1.1689287968911344 0.90632989536970843 0.56352164200507093 1.7144592562690371 1.7879319349303839 1.0440519200637934 0.18746783607639383 0.62702924245968483 0.74173321062698994 0.90928600262850601 1.0372765515931055 1.3575130361132317 0.88587961927988235 1.195353177841753 0.91549000050872553 0.39518852508626923 0.58881598897278276 1.7278680396266013 0.63771811034530379 0.86668374016880989 0.95897737122140825 0.73000309383496631 1.3129157207440718 1.1494598045246665 0.77069143997505296
leaf_count=26 18 45 27 17 34 33 32 13 32 35 31 9 22 44 30 30 35 35 27 25 12 35 50 25 31 56 25 91 48 27
internal_value=0 -0.00237171 0.00189181 0.0470089 -0.00381711 -0.0144697 0.00981766 -0.0393611 0.0266479 0.0450561 0.0127859 0.0420885 -0.00784686 -0.0327368 -0.000146467 -0.0468667 0.0297791 -0.04239 0.0591308 0.033632 -0.069878 -0.0272216 -0.0615091 -0.0718511 -0.0323251 -0.0696368 -0.0902688 0.0950042 0.0965059 -0.0993055
internal_weight=0 29.3242 26.5008 2.97664 23.5242 17.4297 7.48195 2.96439 6.09454 4.54644 2.75851 4.51756 1.41189 9.9477 3.00853 6.93917 2.2668 2.82338 3.61123 2.11084 2.40087 0.984005 5.77024 4.73296 1.5044 1.84486 3.22856 1.50038 1.56475 2.49856
internal_count=1000 982 865 114 751 619 289 105 132 98 63 184 49 330 109 221 65 117 152 52 92 47 188 158 56 91 102 100 65 77
is_linear=0
shrinkage=0.1
Tree=58
num_leaves=31
num_cat=0
split_feature=6 0 15 8 8 3 10 4 9 11 13 2 9 7 26 2 26 24 15 1 11 9 5 16 0 11 9 11 5 14
split_gain=0.430539 0.615949 0.854602 0.543164 0.471989 0.814568 0.582266 0.559656 0.854821 0.783723 0.554514 0.551243 0.51877 0.62468 0.50981 0.471287 0.619072 0.587137 0.479322 0.411416 0.726738 0.502617 0.433753 0.390389 0.372331 0.529465 0.322932 0.263968 0.369457 0.247546
threshold=2.2315721511840825 0.46007883548736578 -0.8999110460281371 0.11012277007102968 1.0175021886825564 -2.7369303703308101 2.1139192581176762 0.4314530491828919 -1.3711115121841428 -0.71831554174423207 -0.44223669171333307 2.3340952396392827 1.444472134113312 0.46503198146820074 -0.43240021169185633 3.0034747123718266 -0.84085938334465016 0.56038278341293346 0.040439540520310409 0.94988065958023082 0.15818990021944049 0.023247927427291874 -0.63844007253646839 -0.4076728075742721 1.3665547370910647 1.9994604587554934 -1.1903105974197385 2.4863992929458623 0.64191484451293956 -0.28190730512142176
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=4 -2 -3 -4 5 -1 15 9 -9 11 27 12 13 23 -12 16 -6 19 -19 20 21 -18 -22 -7 -11 -26 -25 28 -10 -16
right_child=1 2 3 -5 6 7 -8 8 10 24 14 -13 -14 -15 29 -17 17 18 -20 -21 22 -23 -24 26 25 -27 -28 -29 -30 -31
leaf_value=-0.084102377887157687 -0.092784895115051919 -0.10296323758527855 0.096912432384165539 -0.046226283284786382 -0.084540120356485671 -0.093110339172061685 -0.091249542146043028 0.10738505018719545 -0.054265403380913527 0.10198033302550141 0.0061449588816904471 -0.067096856160331106 0.090722524415884936 0.088344200627332781 -0.025975468734724075 0.062482986315711475 -0.10034434594848543 -0.012854487966937595 0.08081541550792154 -0.079253458409445365 -0.0095003405494018187 0.0052755993554905005 0.084152008859661009 -0.039942863473347051 -0.029223156824992647 0.085468075189544357 0.051431654065516447 0.057313785049024846 0.047548922858543868 -0.092861144417061822
leaf_weight=0.79943641507998719 1.1700170631520443 0.54114359943196166 0.78670610650442518 0.39984349999576807 1.064580065663902 0.57048427360132858 0.87513892957940775 0.66830417676828791 0.74363512033596657 1.9467935229185944 1.7410390442237256 0.77343927859328676 0.92231598310172547 0.67061497399117787 0.93325849901884783 0.90855449065566141 1.0037678051739898 0.97127706650644574 1.2485334649682043 1.0049033751711247 1.0517858215607701 0.81749438727274526 0.93344545178115357 0.8108583316206982 0.69237482862081356 0.96143993269652128 0.73953330423682451 1.2610937450081108 0.68444706872105598 1.3592441212385886
leaf_count=31 55 11 47 13 37 19 54 24 26 87 39 29 43 65 29 15 35 27 21 35 29 29 15 18 27 31 30 23 19 37
internal_value=0 -0.0367599 0.00118101 0.0486775 0.00386501 0.0143297 -0.0133776 0.0194134 -0.000477496 0.0375906 -0.0112001 0.00970615 0.0257012 0.00421795 -0.0346505 -0.00580917 -0.0134733 -0.00271323 0.0398302 -0.0223413 -0.00731662 -0.0529356 0.0345345 -0.0223825 0.0723416 0.0374522 0.00364257 0.0239736 -0.00546813 -0.0656325
internal_weight=0 2.89771 1.72769 1.18655 26.1578 16.2783 9.87948 15.4789 7.39102 8.08785 6.72272 4.48725 3.71381 2.79149 4.03354 9.00434 8.09579 7.03121 2.21981 4.8114 3.80649 1.82126 1.98523 2.12088 3.60061 1.65381 1.55039 2.68918 1.42808 2.2925
internal_count=1000 126 71 60 874 577 297 546 197 349 173 204 175 132 105 243 228 191 48 143 108 64 44 67 145 58 48 68 45 66
is_linear=0
shrinkage=0.1
Tree=59
num_leaves=31
num_cat=0
split_feature=0 13 7 3 7 8 6 3 10 4 11 8 22 1 6 9 4 19 27 13 5 8 10 13 14 13 25 8 2 2
split_gain=0.425328 0.445206 0.520478 0.57869 0.516664 0.968159 1.49116 0.659369 0.546385 0.667085 0.980733 0.712814 0.511642 0.506109 0.483451 0.629294 0.610375 0.462327 0.415087 0.403604 0.330036 0.405013 0.263799 0.306935 0.257425 0.225459 0.116397 0.0943502 0.0885586 0.0515606
threshold=-3.1315989494323726 1.4291877746582033 -1.6642212867736814 -0.49797442555427546 -0.20904585719108579 0.72140043973922741 -1.1788545250892637 -1.9173253774642942 -4.9047172069549552 3.1555353403091435 4.503108263015748 -1.8626908063888548 0.89936009049415599 -1.5054519176483152 -1.8197036385536192 0.39663147926330572 -0.31387105584144587 1.5189371109008791 -0.58443316817283619 -0.45099973678588862 -0.7574045956134795 -0.52487343549728382 -1.5154966711997984 0.054351167753338821 -0.15193232148885724 -0.42357698082923884 -0.023944747634232041 1.3702826499938967 0.039535939693450935 1.4999915361404421
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=28 4 -3 -4 7 17 -7 25 -9 10 11 24 14 -8 -13 22 -17 20 -14 -18 21 -6 -16 -24 -10 -2 -12 -15 -1 -27
right_child=1 2 3 -5 5 6 13 8 9 -11 26 12 18 27 15 16 19 -19 -20 -21 -22 -23 23 -25 -26 29 -28 -29 -30 -31
leaf_value=-0.028202292893783228 -0.012088961113004018 0.081651519970688924 -0.067505973668806105 0.053464985328124881 0.092205880493574768 0.06320422558844277 -0.00090974735111622532 -0.075845179972056101 0.102784777624617 -0.069703562011731529 0.046258516057751407 -0.093928675909951714 0.084592111174284529 -0.050742177237205199 0.0052576825037848789 0.078491896663089492 0.03283163910386231 -0.048904143582365972 -0.008537701498479754 -0.060357975567037675 0.097299211275355207 -0.015358228475873887 -0.10548605751969609 -0.027457951640170255 0.020217772948240183 -0.063642210770712429 0.096941198043688059 -0.10272352425723569 -0.10277518401019224 -0.10416261015467197
leaf_weight=0.24363794014789419 0.5128936953842641 1.4346003485843541 0.72311973897740245 0.87266578851267584 0.63325628766324371 1.6214046822860835 1.0214124657213688 1.0295173306949514 0.69387951167300344 1.2489339457824833 0.81722078984603275 1.0037855524569774 0.78132970537990365 0.47397293290123355 0.58829825418070303 0.97057393519207869 0.83360770717263233 0.37200919957831491 1.2351633752696214 1.050324982963502 1.7972734103677774 0.78273736964911289 1.1323148910887306 0.90871217986568797 0.82843473763205533 0.48648927873001435 1.0170778315514315 1.326194714521989 0.45974109275266528 0.88584006344899535
leaf_count=13 18 28 26 29 46 39 31 25 25 28 23 38 18 13 17 34 16 17 35 25 146 23 36 34 38 26 21 64 35 33
internal_value=0 0.0018208 0.0379422 -0.00135221 -0.00273012 0.0179762 -0.0132189 -0.013104 -0.00569703 -0.000188206 0.00713176 -0.00516758 -0.0164487 -0.057135 -0.0301236 -0.0184445 0.0140674 0.0566341 0.0275473 -0.0191232 0.0688526 0.0327463 -0.0537407 -0.0707462 0.0578523 -0.0686566 0.0743609 -0.0890372 -0.0769445 -0.0897982
internal_weight=0 27.083 3.03039 1.59579 24.0527 8.02826 4.44298 16.0244 14.1392 13.1097 11.8607 10.0264 8.50411 2.82158 6.48762 5.48383 2.85451 3.58528 2.01649 1.88393 3.21327 1.41599 2.62933 2.04103 1.52231 1.88522 1.8343 1.80017 0.703379 1.37233
internal_count=1000 952 83 55 869 379 147 490 413 388 360 316 253 108 200 162 75 232 53 41 215 69 87 70 63 77 44 77 48 59
is_linear=0
shrinkage=0.1
Tree=60
num_leaves=31
num_cat=0
split_feature=20 8 4 11 9 8 7 6 4 31 30 2 12 9 24 1 11 2 5 2 6 21 4 23 20 8 0 4 2 5
split_gain=0.423866 0.43037 0.609234 0.635065 0.851823 0.790379 0.764102 0.744749 0.637476 0.613895 0.597059 0.581336 0.478103 0.428362 0.414554 0.522311 0.549792 0.413475 0.402334 0.398418 0.388089 0.334144 0.296904 0.242454 0.235647 0.231239 0.244342 0.171969 0.149035 0.115038
threshold=1.9680437445640566 2.9477462768554692 -3.4333997964859004 -0.97970148921012867 -1.9159966707229612 -1.7129650115966795 -0.88453030586242665 2.2315721511840825 -1.6822952032089231 1.3691098093986513 0.45831295847892767 -1.4813745021820066 0.36355853080749517 -1.1903105974197385 -1.1339773535728452 1.0527601242065432 2.4600665569305424 -1.6195735931396482 -0.67728805541992176 1.8555500507354739 0.32099404931068426 0.83956339955329906 1.5504462718963625 -0.66370597481727589 -0.37434956431388849 1.3702826499938967 -1.608339190483093 -0.74094659090042103 0.29434275627136236 -0.7574045956134795
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 2 13 6 21 22 11 8 -7 12 17 -4 14 -1 -10 16 18 -8 -16 -3 -17 27 -6 -14 -12 26 -19 -5 -13 -25
right_child=-2 19 3 4 5 7 10 -9 9 -11 24 28 23 -15 15 20 -18 25 -20 -21 -22 -23 -24 29 -26 -27 -28 -29 -30 -31
leaf_value=0.0029138584466358525 0.069439738688708841 -0.081383221892300878 0.044169620375901546 0.034258455737236862 0.10303647431628174 0.057603926001794176 -0.069686767938545141 -0.089990771483029211 -0.063453652835031168 0.061181777791627062 0.015508936523029544 -0.051071740695264024 -0.013681752676736226 0.10410224247437812 -0.07313597830797991 0.010191033036591033 0.072681370709638621 0.0021503195415847067 0.038079753141781414 0.016851945965725017 -0.091542570461229711 0.00091407041040665604 0.02016218616031068 -0.10184043505822443 -0.085163443042067688 -0.013811933676917782 0.093365115776898031 0.10258802297377959 -0.099132017329310165 -0.048412517882624677
leaf_weight=0.75525410985574126 0.8467904687859108 1.2437229824718077 0.42665383778512511 0.51178141636773977 0.8422923011239678 1.4582484220154643 0.33359025581739898 1.1064295694231976 1.0264815930277151 0.97244759742170561 0.32798327808268368 1.0275197378359737 0.79054266354069147 0.93788480688818754 0.62771325372159836 0.75163727998733498 1.5457486039958892 0.39915629918687079 0.67512175440784117 0.61801664857193817 0.74827108811587095 0.67103576497174788 0.88808668311685324 1.0228085448034108 0.79874956095591165 0.43060001078993093 1.1113168291049169 1.3140094750560916 1.7342747463844714 0.66503054159693387
leaf_count=33 18 45 15 29 62 35 14 43 23 25 23 37 30 65 33 20 38 19 17 19 27 16 26 37 38 15 58 43 77 20
internal_value=0 -0.00246359 0.00114371 -0.00326471 0.00771998 -0.00246779 -0.029298 -0.0120339 -0.00364678 -0.013767 0.00367046 -0.0644684 -0.0230476 0.0589654 -0.00629315 0.00719985 0.0323484 0.0331576 -0.0155046 -0.0487734 -0.0405616 0.0612569 0.0605027 -0.0593835 -0.0558585 0.0508323 0.0692608 0.0834348 -0.0812513 -0.0807891
internal_weight=0 25.7624 23.9007 22.2075 15.6177 13.1209 6.58984 11.3905 10.2841 8.8258 3.4014 3.18845 7.85336 1.69314 5.37497 4.34849 2.84858 2.27466 1.30284 1.86174 1.49991 2.49683 1.73038 2.47838 1.12673 1.94107 1.51047 1.82579 2.76179 1.68784
internal_count=1000 982 918 820 524 436 296 348 305 270 167 129 245 98 158 135 88 106 50 64 47 88 88 87 61 92 77 72 114 57
is_linear=0
shrinkage=0.1
Tree=61
num_leaves=31
num_cat=0
split_feature=6 0 2 6 20 8 3 2 9 13 9 10 15 24 1 9 27 8 9 23 6 28 6 17 4 24 0 26 8 26
split_gain=0.385867 0.57805 0.635524 0.490261 0.392927 0.414558 0.559544 0.581665 0.824828 0.751173 0.547645 0.530256 0.522418 0.430445 0.58094 0.444051 0.405896 0.38869 0.386509 0.368038 0.343267 0.309919 0.299399 0.296172 0.293248 0.269231 0.219735 0.2133 0.12596 0.0922573
threshold=1.7042843699455263 0.06608057022094728 1.4812332987785342 1.5142216682434084 1.9680437445640566 1.4360643625259402 0.12777689099311831 0.58200773596763622 -1.0644020438194273 -0.18393839150667188 0.82368111610412609 -3.4719688892364498 0.074047245085239424 -0.46206876635551447 1.2805632352828982 1.2716721296310427 -0.30689744651317591 -0.75641953945159901 -2.3101347684860225 -0.52800947427749623 -1.5804541110992429 0.5733412802219392 -0.60408353805541981 -0.38247415423393244 -1.4721328616142271 0.91924515366554271 -0.641687512397766 0.16083426773548129 -0.45424538850784296 -0.84867754578590382
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=3 29 24 4 5 6 10 18 -9 22 11 19 -11 15 20 -7 -12 -4 -8 -1 -15 -13 -10 -22 -3 -20 -18 -25 -24 -2
right_child=1 2 17 -5 -6 13 7 8 9 12 16 21 -14 14 -16 -17 26 -19 25 -21 23 -23 28 27 -26 -27 -28 -29 -30 -31
leaf_value=0.075972430803731908 -0.032971298607464943 0.0075464312198496994 0.07992234727340114 0.076408579380212574 0.071642845412225792 -0.096616533121099954 -0.048149415521922458 0.10484750741605041 -0.018136293648719944 0.0088754704553617135 -0.024070232577895031 -0.083097907071876462 -0.0832776554687992 -0.03415508833138764 -0.089789175183350389 0.019639606464482373 0.014456768196181448 -0.0076962260535432736 0.098882157257243045 -0.028326352881101902 0.080342869742464809 0.0018914262211941969 0.03795473589071837 -0.033427471485797507 -0.092612797739701019 0.011287027960921124 0.098363648357424893 0.045928101898803038 0.10585269429406743 -0.094780435922102182
leaf_weight=0.55848215706646442 0.31289158295840014 0.44433678360655904 0.90747343911789391 0.93613336468115349 0.78602411691099305 1.2328037805855272 0.25815017218701575 0.74225426279009266 0.53342253272421647 1.0742375759873368 0.92272976646199856 1.9091156830545641 1.4395432788878677 1.0821257818024608 0.6751344222575425 0.44792447402141977 0.54011130426079046 1.1452928120270369 1.7428733513224872 0.8582415394484999 0.86280746664851982 0.55344368051737536 0.52587845595553506 0.69240964716300257 0.85440300730988405 0.4393353642662986 0.73933776700869169 0.66309204162098467 0.56869181431830762 1.0582019363064308
leaf_count=14 19 32 49 32 17 54 11 41 28 39 35 75 40 36 29 20 22 37 88 34 19 23 26 26 28 19 22 17 15 53
internal_value=0 -0.0259763 -0.00359913 0.00568937 0.00235392 -0.000503195 0.00907593 0.027692 0.00777205 -0.00962501 -0.0133449 -0.0359547 -0.043897 -0.0232064 -0.00526978 -0.0656336 0.0264837 0.0310376 0.067559 0.0127889 0.0120194 -0.0639971 0.0432943 0.0345441 -0.0583454 0.081247 0.0629429 0.00539214 0.0732316 -0.0806752
internal_weight=0 4.7226 3.35151 20.7843 19.8482 19.0621 13.4058 7.32439 4.88403 4.14177 6.08146 3.87928 2.51378 5.6563 3.97557 1.68073 2.20218 2.05277 2.44036 1.41672 3.30043 2.46256 1.62799 2.21831 1.29874 2.18221 1.27945 1.3555 1.09457 1.37109
internal_count=1000 218 146 782 750 733 532 307 189 148 225 146 79 201 127 74 79 86 118 48 98 98 69 62 60 107 44 43 41 72
is_linear=0
shrinkage=0.1
Tree=62
num_leaves=31
num_cat=0
split_feature=0 1 8 23 20 1 6 6 7 21 1 5 2 6 1 9 1 7 3 7 9 31 24 22 24 8 20 5 3 4
split_gain=0.390649 0.467561 0.520943 0.650594 0.450959 0.657389 0.737523 0.59653 0.547421 0.527418 0.402475 0.462741 0.3903 0.515533 0.520205 0.370639 0.355209 0.408467 0.339429 0.326358 0.320967 0.308218 0.350466 0.372822 0.29824 0.266295 0.251974 0.212446 0.143913 0.106173
threshold=-1.608339190483093 3.992965936660767 -1.9779059886932371 2.3457252979278569 0.45991475880146032 0.81392666697502147 -1.5804541110992429 -1.5804541110992429 -0.43933647871017451 0.034177251160144813 1.3164123892784121 0.27409940958023077 2.1508438587188725 -0.27595347166061396 -1.6003860235214231 0.42599982023239141 -0.93864721059799183 -1.7502976059913633 0.18194845318794253 -0.20904585719108579 -1.9159966707229612 0.88406136631965648 -0.30356071889400477 -0.019407821819186207 0.12283229455351831 0.93145474791526806 -0.58664342761039723 -0.89550951123237599 -0.053450405597686761 1.7364857196807864
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=10 2 3 29 5 7 -7 8 -4 16 11 -1 13 25 -15 20 26 -18 -8 -19 -16 22 27 -24 -20 -6 -9 -11 -14 -2
right_child=1 -3 4 -5 12 6 18 9 -10 21 -12 -13 28 14 15 -17 17 19 24 -21 -22 -23 23 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.089787344331220334 0.099242332197471916 -0.073673939103793995 -0.091019048664033528 -0.087391885436217773 0.080001783821247396 0.045351610820920513 -0.095000927846191977 0.10808542640810266 0.02686833754679049 -0.10578868539565828 0.04041741642425703 0.0010426023470820561 0.037577428830497271 0.070922108057272101 -0.0052696603913138249 0.027768407236546673 0.068039393038333257 -0.090149103402227593 -0.065136613259972748 0.031358691486546596 -0.1044492358421133 0.052213001534613258 0.06908338671754409 -0.033398236180782789 0.032034404032421769 0.013653367898637149 0.038931166151211172 -0.022416694763460875 0.097106526995900264 0.040224104909689001
leaf_weight=1.387808014173064 1.1263660714030326 0.73934661969542403 1.0566249974071977 0.26114840665832151 1.0497015428263696 0.74874908197671208 1.7054309258237474 1.0124180046841504 0.62802197854035391 0.6103771859779954 0.57386932242661703 0.94134266907349218 0.8480844465084374 0.47591747855767619 0.49118768819607772 0.51102546206675481 0.67318714247085176 0.48826688667759321 0.81865537725388982 0.40390192810445336 0.97201660415157587 0.65060518658719946 0.59975147410296681 0.86981927440500217 0.51428204553667456 1.4276593774557111 1.0986528429202738 0.61218138865661675 0.77926249080337562 0.41791638452559698
leaf_count=95 132 24 38 4 40 33 72 21 23 18 22 35 25 12 23 23 18 17 26 25 31 26 23 32 27 38 29 25 29 14
internal_value=0 0.00447607 0.00724712 0.058585 0.00238058 -0.0087663 -0.0435451 0.0063663 -0.0470716 0.0191918 -0.0345956 -0.0530778 0.0236221 0.00959897 -0.0229257 -0.0455491 0.0453298 0.00923257 -0.065452 -0.0351402 -0.0711553 -0.00955553 -0.0244831 0.00842588 -0.0276455 0.0417664 0.0720959 -0.0640412 0.0660832 0.0832707
internal_weight=0 21.5906 20.8512 1.80543 19.0458 12.4909 3.78712 8.70381 1.68465 7.01916 2.90302 2.32915 6.55486 4.92751 2.45015 1.97423 3.67643 1.56536 3.03837 0.892169 1.4632 3.34273 2.69213 1.46957 1.33294 2.47736 2.11107 1.22256 1.62735 1.54428
internal_count=1000 848 824 150 674 453 158 295 61 234 152 130 221 167 89 77 110 60 125 42 54 124 98 55 53 78 50 43 54 146
is_linear=0
shrinkage=0.1
Tree=63
num_leaves=31
num_cat=0
split_feature=0 1 8 7 13 6 29 24 3 0 1 4 8 11 1 19 6 7 7 5 29 9 7 8 6 8 19 2 2 26
split_gain=0.336722 0.34652 0.363451 0.774829 0.432006 0.406931 0.446155 0.376852 0.379756 0.42469 0.508647 0.573702 0.461612 0.391884 0.368568 0.35086 0.326952 0.625927 0.556134 0.453026 0.398475 0.556179 0.320545 0.306174 0.357453 0.267866 0.249003 0.166059 0.0708568 0.0696202
threshold=-3.1315989494323726 4.1692261695861825 0.62742966413497936 -0.20904585719108579 1.4291877746582033 0.40676501393318182 -0.4948394894599914 1.6863377690315249 -2.7369303703308101 -1.349173307418823 -2.2575167417526241 0.90728291869163524 -1.5996349453926084 1.0588229894638064 -0.051437556743621819 -0.016774930991232392 -0.41582041978836054 0.25074130296707159 -0.25185197591781611 -0.80016306042671193 0.5306413769721986 1.2301043272018435 -1.1648204326629636 -0.52487343549728382 0.79199358820915233 -1.0830669999122617 0.73767346143722545 0.68040439486503612 0.023640871047973636 -0.64738211035728443
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=28 2 3 4 7 16 -7 8 -2 25 -11 12 -12 14 -14 -13 17 20 -18 -19 21 -4 -22 -5 27 -10 29 -25 -1 -8
right_child=1 -3 5 23 -6 6 26 -9 9 10 11 15 13 -15 -16 -17 18 19 -20 -21 22 -23 -24 24 -26 -27 -28 -29 -30 -31
leaf_value=-0.028647310603651839 -0.10247567684083461 -0.073619422533753892 -0.10434422749257416 0.096382356212668469 0.066094554615706874 0.01226099480475199 -0.052637829799616835 0.077150646791427038 -0.01614943791903339 0.10316567950811831 0.10216608919366976 -0.079095731242918854 0.023500694295390964 0.045760378823160239 -0.092588969763089099 -0.0010785981842537284 0.080586495716705203 -0.032153928889815983 -0.024072827870040017 0.074055371708330486 -0.044742521499505875 -0.0014513937466770323 0.051127288992983945 0.018180114456665945 -0.028468539048135417 -0.10357276851456626 -0.011591471991762345 0.1019976535059022 -0.10215104867513698 -0.10338517900721472
leaf_weight=0.20371210831217723 0.43611589772627013 0.59474529791623254 1.2620040443725891 1.3092372694518419 0.91877906769514162 1.0405288725160065 0.36160148214548815 0.50719480216503132 0.78236777521669865 0.49093749350868421 0.62868974031880798 1.1656631743535402 0.51976181252394171 1.1616725549101827 0.57717467565088387 1.1403702767565844 1.2012918773107233 0.6087808497250079 0.87938860058784296 1.1800988893955944 0.64725709799677134 0.89999010087922204 0.75624364847317316 0.39318250794894882 0.61732665961608291 0.6348962853662703 0.55323985754512239 0.59265683684498072 0.36817994923330843 1.0711366455070672
leaf_count=12 26 20 44 161 28 37 16 16 24 20 54 28 25 34 23 27 31 26 29 30 21 34 19 23 23 29 19 33 36 52
internal_value=0 0.00171484 0.00372064 0.0156927 0.00113299 -0.00987007 -0.0407825 -0.00628609 -0.0119004 -0.00633803 0.00587288 -0.00332443 0.026379 0.00528349 -0.0375822 -0.040515 0.00271312 -0.0103591 0.0363528 0.0379109 -0.0345771 -0.0615122 0.00691469 0.0605035 0.0312028 -0.0553127 -0.068574 0.0685687 -0.0759685 -0.0905773
internal_weight=0 22.9323 22.3376 11.876 8.96362 10.4616 3.02651 8.04484 7.53765 7.10153 5.68427 5.19333 2.8873 2.25861 1.09694 2.30603 7.43506 5.35437 2.08068 1.78888 3.56549 2.16199 1.4035 2.9124 1.60317 1.41726 1.98598 0.985839 0.571892 1.43274
internal_count=1000 952 932 574 334 358 124 306 290 264 211 191 136 82 48 55 234 174 60 56 118 78 40 240 79 53 87 56 48 68
is_linear=0
shrinkage=0.1
Tree=64
num_leaves=31
num_cat=0
split_feature=30 30 0 8 1 20 1 27 11 4 9 1 4 8 10 21 8 18 10 5 5 17 31 13 31 13 21 10 5 16
split_gain=0.3142 0.446132 0.422177 0.517312 0.419013 0.445745 0.461262 0.542245 0.379104 0.479529 0.436849 0.366441 0.352366 0.465019 0.349338 0.407973 0.504054 0.37696 0.334953 0.320853 0.257895 0.270018 0.309709 0.245767 0.354995 0.229561 0.222406 0.167864 0.154037 0.144112
threshold=-1.2419461011886594 -1.1328976750373838 -1.608339190483093 -2.5952844619750972 -3.7169742584228511 0.35997511446475988 0.77400022745132457 0.60264223814010631 0.77870225906372081 -1.695614159107208 -1.4426879882812498 1.4367862939834597 1.4679937958717348 0.84256586432456981 -4.4069178104400626 -0.4803245067596435 0.62742966413497936 -0.65785545110702504 -2.5471062660217281 0.33765733242034918 -2.186565518379211 -0.69839975237846363 0.096631966531276717 -0.65081390738487233 0.17010550200939181 0.52293393015861522 -1.0671396255493162 2.1767039299011235 1.4445725679397585 -1.1121394634246824
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=12 -2 11 29 -5 6 14 26 9 10 -7 19 13 -1 -6 23 17 -17 -18 -3 -10 -22 -23 -16 -25 -20 -8 -28 -11 -4
right_child=1 2 3 4 5 8 7 -9 20 28 -12 -13 -14 -15 15 16 18 -19 25 -21 21 22 -24 24 -26 -27 27 -29 -30 -31
leaf_value=0.052687374003742349 0.083691980891362927 -0.093919641941266993 -0.0051087952765356097 0.071189466828763842 -0.065048233634231067 -0.023642694410921601 0.0016065893765541709 0.034331850260422729 -0.012500453220566921 -0.085685864901957606 0.092700136515807829 0.035177925363860676 -0.093396040614032685 -0.069523604561661789 0.089563023303226186 -0.027989867897796596 0.020402421501915506 0.068167534642200642 -0.096629170696494418 -0.009968505104585643 0.10303413854280903 0.073699146359801362 -0.010332485273629819 0.066842017777253612 -0.044334664710113061 -0.016175263445588912 -0.0878022579044649 -0.020658713650305124 -0.0097425560714198556 0.10229713682517663
leaf_weight=0.64896089653484768 0.67081491055432962 1.1007886860752578 0.15269144391641021 0.83881153259427677 0.69773427373730612 0.66809541697148234 0.45318859897088337 0.72408303641714145 0.6661341008730236 0.69772834354080149 0.62434238207060866 0.52687265211716283 0.7306126826442777 0.59848472103476524 0.8008663214277475 0.66213012114168224 0.62139423750340972 1.0609296433394213 0.95490834489464738 0.77631161827594031 0.86211982858367342 0.94623266370035641 0.81755602860357612 0.64829935505987168 0.51564632705412805 0.5641972916200757 1.9244486453244465 0.46167288220021863 0.43272539996542037 0.6869393435772605
leaf_count=45 24 88 12 27 29 27 21 36 30 35 35 20 23 20 17 25 14 54 40 31 20 27 32 24 21 17 82 25 22 77
internal_value=0 0.00348234 0.00077674 0.00617959 0.00231596 -0.00133946 -0.0139788 -0.0429145 0.0209748 -0.00904715 0.0325594 -0.0385154 -0.0382461 -0.00594557 0.00182072 0.00982583 -0.00904142 0.0312166 -0.0414483 -0.0592 0.0430705 0.0571676 0.0347486 0.0469259 0.017589 -0.0667485 -0.0626139 -0.0748112 -0.0566156 0.0827648
internal_weight=0 20.5577 19.8868 17.4829 16.6432 15.8044 10.0895 3.56339 5.71493 2.42289 1.29244 2.40397 1.97806 1.24745 6.52611 5.82837 3.86356 1.72306 2.1405 1.8771 3.29204 2.62591 1.76379 1.96481 1.16395 1.51911 2.83931 2.38612 1.13045 0.839631
internal_count=1000 912 888 749 660 633 405 164 228 119 62 139 88 65 241 212 150 79 71 119 109 79 59 62 45 57 128 107 57 89
is_linear=0
shrinkage=0.1
Tree=65
num_leaves=31
num_cat=0
split_feature=6 0 2 6 25 25 10 12 5 6 6 1 0 7 29 25 17 7 10 16 0 4 4 31 4 18 18 19 7 29
split_gain=0.305794 0.455563 0.473317 0.369166 0.353033 0.398957 0.534566 0.36583 0.321552 0.770537 0.594101 0.541365 0.422077 0.420729 0.452802 0.323188 0.305308 0.301188 0.298577 0.294551 0.254988 0.317977 0.22617 0.189304 0.254218 0.186476 0.30487 0.313563 0.18825 0.1744
threshold=1.7042843699455263 0.06608057022094728 1.4812332987785342 1.5142216682434084 0.36171150207519537 0.31395921111106878 -5.7338836193084708 -0.82562109827995289 -0.51409405469894398 -1.5258474349975584 -0.19717428088188169 0.35094490647315985 2.5667488574981694 -1.5177835822105405 0.48870287835597997 -0.63787114620208729 0.56251215934753429 -2.3474750518798824 1.4502543807029726 0.93398618698120128 1.3665547370910647 -1.1332715749740598 -1.4721328616142271 -0.20878539979457852 1.0000000180025095e-35 -0.58789622783660878 0.13855654746294024 0.13681121915578845 0.51377156376838695 -0.20581085234880445
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=3 -2 22 4 5 6 -1 20 9 -8 17 15 13 18 29 -12 -4 -10 -11 25 21 -6 -3 24 -19 -9 27 -27 -28 -15
right_child=1 2 16 -5 7 -7 8 19 10 12 11 -13 -14 14 -16 -17 -18 23 -20 -21 -22 -23 -24 -25 -26 26 28 -29 -30 -31
leaf_value=-0.06281393488840023 -0.078704094822772463 0.005604328586266109 0.060149821979332788 0.072291143946570921 0.061527867943828134 0.098477882952643078 -0.099142757116850777 -0.091419887695948457 0.004174386100491671 0.090085571560383285 0.089484376101165039 -0.088520979607655051 -0.057816124143210683 -0.01070760134794657 0.048981640166312741 -0.022587580668022092 -0.027098123607685012 0.00054212956991272424 0.014179191300432709 0.024966892829264859 0.07399736219176091 -0.057668723520005682 -0.088428193874684813 0.10318001757516909 0.10090142315623447 0.07179738481581624 -0.085267408742940476 -0.051275230517136228 -0.0069362051504830912 -0.089343296355734014
leaf_weight=0.97059407201596037 1.1369935187976801 0.3983466110075824 1.1134602499660102 0.78986255079507728 0.37179491901770234 0.51071930280886668 0.70573635818436731 0.77467593207256902 0.72959794406779543 1.1680647074244914 0.48421138618141402 0.61579693062230934 0.7178388126194476 0.57741956261452232 0.80687953322194506 0.54911605129017749 0.62688524019904424 0.56203196558636126 0.93142199795693148 0.65296907210722555 0.58440891583450139 0.5622579902410505 0.71473157755099259 1.2979662405559749 0.45814952463842928 0.438778064562939 1.0349875458050517 0.39192469802219421 0.43607588973827649 0.55133167561144159
leaf_count=34 72 32 58 32 17 19 45 38 27 26 21 35 28 30 31 25 28 37 30 38 20 26 28 60 15 28 49 21 25 25
internal_value=0 -0.0251776 -0.00384904 0.00547099 0.00234517 0.0120555 0.00808849 -0.019187 0.0148648 -0.00164087 0.0340477 -0.0143014 0.0128366 0.0254055 -0.00822382 0.0299287 0.0287225 0.0602092 0.0564103 -0.0360342 0.0221907 -0.0102231 -0.054776 0.0778453 0.0456121 -0.0489816 -0.0346987 0.0137319 -0.0620472 -0.0491167
internal_weight=0 3.99042 2.85342 17.6746 16.8847 11.6369 11.1262 5.24787 10.1556 5.45869 4.69687 1.64912 4.75296 4.03512 1.93563 1.03333 1.74035 3.04775 2.09949 3.72941 1.51846 0.934053 1.11308 2.31815 1.02018 3.07644 2.30177 0.830703 1.47106 1.12875
internal_count=1000 218 146 782 750 488 469 262 435 215 220 81 170 142 86 46 86 139 56 199 63 43 60 112 52 161 123 49 74 55
is_linear=0
shrinkage=0.1
Tree=66
num_leaves=31
num_cat=0
split_feature=3 8 4 11 1 7 4 8 6 11 30 5 1 6 18 18 8 6 2 23 10 5 31 12 20 9 8 4 23 14
split_gain=0.309954 0.389055 0.426778 0.49406 0.442363 0.503347 0.40649 0.378606 0.58212 0.372228 0.382854 0.353829 0.456437 0.389409 0.511571 0.35319 0.339382 0.368401 0.338694 0.330702 0.369196 0.328362 0.262201 0.336163 0.330401 0.239294 0.175428 0.199805 0.280674 0.201202
threshold=-1.9173253774642942 2.9477462768554692 -3.4333997964859004 -0.97970148921012867 -0.91019541025161732 -0.88453030586242665 1.7364857196807864 -1.9779059886932371 0.77008655667305004 3.7105647325515752 0.99595102667808544 -0.51409405469894398 -1.1366724371910093 -0.59109562635421742 0.11296719312667848 -0.25806640088558191 1.0356987714767458 2.2315721511840825 -0.33402413129806513 0.32948938012123113 -1.5789780020713804 -0.63844007253646839 1.3691098093986513 -0.53976348042488087 0.089713446795940413 -1.1903105974197385 0.071703881025314345 -0.12080970406532286 0.77710625529289257 -0.26860181987285608
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=19 2 25 4 5 18 7 -5 9 10 22 13 -13 -8 -15 -10 17 -17 -4 20 -1 -3 23 -9 -25 -2 27 28 29 -6
right_child=1 21 3 6 26 -7 11 8 15 -11 -12 12 -14 14 -16 16 -18 -19 -20 -21 -22 -23 -24 24 -26 -27 -28 -29 -30 -31
leaf_value=0.032587806524009277 0.019201332913234087 -0.10268365264063846 0.029329528527959621 0.10210723351060456 0.098495889565655761 0.06409666389812689 -0.10391510445074266 0.072181338281376226 -0.072967985731719684 0.095891832919396888 -0.051913022867497438 0.091551349946131913 -0.023586645288458212 0.06314623243364581 -0.062650416320150809 0.0826668051497223 -0.065701026711528429 -0.0285878898477986 -0.081268092438490411 -0.084532141901636035 -0.062905781349168341 -0.0068428935780731354 0.10290828122675576 -0.053697086970651113 0.042339594939967362 0.10300577345404255 -0.086266311856754999 -0.0890587678409374 -0.075119398779260216 -0.016905422044725865
leaf_weight=0.96648286911658954 0.66215140884742141 0.58841922390274704 0.47581113711930811 0.63496775610838518 0.27619328408036392 0.93142568820621807 0.65347351902164508 0.97307373723015733 0.78627929463982615 1.0212807361967859 0.51989305019378651 0.59361765161156654 0.81979678454808869 0.55907911341637395 0.76644912525080144 0.70531377801671602 0.45222854614257801 0.51493085455149412 0.66234293836168967 0.83537172712385643 0.69672422017902413 0.91084045730531105 0.55764002818614233 0.69172110734505488 0.74305426306091249 0.7018871525651853 0.97410863894037891 0.41549873910844959 0.36908660759218037 0.33352069382090122
leaf_count=41 33 37 18 67 27 54 33 29 39 30 20 23 27 16 24 32 24 17 30 49 34 26 14 35 30 62 62 24 16 27
internal_value=0 0.00435115 0.00870829 0.00396879 -0.0241928 0.00958129 0.0153381 0.0281853 0.0214466 0.0427997 0.0272428 -0.0134447 0.02477 -0.0407379 -0.00959211 -0.0176918 0.00829512 0.0357185 -0.0350322 -0.0331981 -0.00741485 -0.0444578 0.0411199 0.0268102 -0.00396075 0.0623242 -0.0537055 -0.0309573 -0.00629339 0.03537
internal_weight=0 18.2941 16.7948 15.4308 4.43799 2.06958 10.9928 7.60038 6.96542 4.50666 3.48538 3.39242 1.41341 1.979 1.32553 2.45875 1.67247 1.22024 1.13815 2.49858 1.66321 1.49926 2.96549 2.40785 1.43478 1.36404 2.36841 1.3943 0.978801 0.609714
internal_count=1000 876 813 718 258 102 460 337 270 158 128 123 50 73 40 112 73 49 48 124 75 63 108 94 65 95 156 94 70 54
is_linear=0
shrinkage=0.1
Tree=67
num_leaves=31
num_cat=0
split_feature=18 18 3 8 4 2 6 22 28 22 30 24 27 23 17 14 8 7 8 12 12 19 8 5 0 10 1 4 24 14
split_gain=0.27674 0.443392 0.339725 0.382116 0.353681 0.385695 0.370156 0.347698 0.362262 0.363126 0.400906 0.423422 0.333456 0.313697 0.284517 0.258936 0.249779 0.239106 0.22184 0.215652 0.203403 0.191791 0.189742 0.188523 0.180671 0.169199 0.167665 0.351941 0.135768 0.133286
threshold=-1.4502596855163572 -0.79747301340103138 -1.9173253774642942 2.9477462768554692 2.8194653987884526 3.5605782270431523 2.1820205450057988 0.88484010100364696 -1.1855872869491575 -0.28644514083862299 0.014556398149579765 -0.64478895068168629 -1.1444159746170042 0.73227164149284374 -0.14468515664339063 -0.64729103446006764 -0.92787450551986683 -2.6797555685043331 0.93145474791526806 -0.13493857532739637 -1.1241716742515562 0.13681121915578845 1.1326235532760622 -0.51409405469894398 0.46007883548736578 -3.8521424531936641 0.81392666697502147 -0.072602182626724229 0.86320912837982189 -1.0000000180025095e-35
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=12 20 16 4 5 6 7 8 29 14 11 -11 -1 -5 21 -9 -3 -18 -14 -13 -2 -10 -17 -6 -8 -19 27 -12 -22 -4
right_child=1 2 3 13 23 -7 24 15 9 10 26 19 18 -15 -16 22 17 25 -20 -21 28 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.04628721264046344 0.018510312784100383 0.031939808755839065 -0.011906678147690074 -0.088579811755687354 -0.069831526622472773 0.085675351273696673 -0.090265853332555185 -0.011470930067781524 -0.015959445817672429 -0.053605724434385882 0.050484473487525432 0.090558942179598811 0.091727249906686847 0.016505026579613863 0.075661194589148162 0.099431455707990321 0.0054083217714528276 -0.020666947457975305 0.012179349734198007 0.013337947351973915 -0.069424613147775094 0.062282514593386032 0.026461914068746518 0.0089898259404044556 -0.00025196522616556852 -0.10275488168030233 -0.08250545014508559 -0.06525761056816784 -0.0026192001555395087 -0.087763478137698483
leaf_weight=0.38096903613768507 0.43430433492176268 0.44116017944179509 0.47477318113669753 0.70460845867637434 0.73051039827987552 0.80529352766461748 0.47106433985754925 0.44299241359112818 0.87526324682403378 0.48033693339675654 0.45853213465306708 0.77851975092198689 0.84884079219773656 0.47596543864347041 1.4518206467619146 1.0614159632241351 0.49794327514246117 0.3640053840354085 0.59724037605337799 0.67537871736566613 1.6319329241523566 0.48794745770283043 0.53646034700795997 0.51904630893841275 0.42340158869046718 0.8094996043946634 0.58134843618609 0.6151890169130656 0.37391215073876072 0.45229128620121439
leaf_count=13 19 26 26 38 27 38 28 22 42 22 33 27 41 14 76 54 18 17 26 38 97 30 18 20 36 42 33 30 20 29
internal_value=0 -0.00389487 0.00230006 0.00813457 0.0133418 0.0190335 0.0138063 0.0196722 0.00950986 0.0179672 -0.00312042 0.0277946 0.0369458 -0.0462133 0.044855 0.0561782 -0.0349896 -0.0526549 0.0588735 0.0546875 -0.0435369 0.0120465 0.0749331 -0.0370904 -0.0476572 -0.0772923 -0.03925 -0.01583 -0.0569713 -0.0489153
internal_weight=0 18.0549 15.6148 13.5022 12.3216 11.072 10.2667 9.37227 7.3314 6.40434 3.5893 1.93424 1.82705 1.18057 2.81503 2.04087 2.11261 1.67145 1.44608 1.4539 2.44015 1.36321 1.59788 1.24956 0.894466 1.1735 1.65507 1.07372 2.00585 0.927064
internal_count=1000 920 784 681 629 582 544 480 386 331 183 87 80 52 148 94 103 77 67 65 136 72 72 47 64 59 96 63 117 55
is_linear=0
shrinkage=0.1
Tree=68
num_leaves=31
num_cat=0
split_feature=5 3 5 1 11 3 19 1 10 8 7 4 9 9 28 8 6 23 8 24 21 31 13 2 26 27 24 19 17 23
split_gain=0.266875 0.298481 0.499435 0.410318 0.358796 0.559498 0.52566 0.475265 0.472245 0.380746 0.553952 0.503686 0.706696 0.371423 0.347342 0.32145 0.375825 0.286626 0.25848 0.212539 0.212283 0.32735 0.206517 0.183368 0.174267 0.35108 0.164452 0.109671 0.0827789 0.04811
threshold=-4.0636310577392569 2.6622943878173833 -0.67728805541992176 2.8490761518478398 2.1326044797897343 1.3817856311798098 0.37843565642833715 0.31563207507133489 0.88901996612548839 2.7939518690109257 1.0094155073165896 -0.22178694605827329 0.12808340787887576 1.2716721296310427 -0.014420712832361458 -0.40550351142883295 0.10272473096847536 -0.5851946473121642 -0.2367703318595886 -0.4543366134166717 -0.49505710601806635 1.2021329402923586 -1.3890017867088316 -0.47287243604660029 0.19097659736871722 0.3827961534261704 -0.49581414461135859 0.19782204926013949 -0.10486884042620657 -0.20161423832178113
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=-1 3 17 4 8 7 -7 26 9 10 11 12 20 18 -15 -9 -17 -3 -13 -5 -2 28 -10 -12 25 29 -6 -14 -22 -24
right_child=1 2 -4 19 5 6 -8 15 22 -11 23 13 27 14 -16 16 -18 -19 -20 -21 21 -23 24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.0747447315319658 0.034638912670394539 0.049860929577762864 0.092098632210348197 0.0040578443008904763 0.03617654100746797 -0.069507424758032621 0.047488170678765096 0.081895231186644851 0.065986065300712091 -0.06527281862863811 0.012843969780741232 -0.016855429309009205 0.045070722812846492 -0.068168556707878455 0.050452441049323372 0.026175035357067555 -0.082156927679088013 -0.056206816173063306 -0.099663764818988351 -0.08430036996363828 -0.037941676856800069 0.028143146815220517 -0.03641852632348571 0.10021658461231764 -0.085517453791906339 0.050464343496153509 0.10012108552353041 0.10321644919950437 -0.10079838284472914 -0.087615423897183811
leaf_weight=0.46768219675869005 0.60449377750047006 0.458205197006464 0.92405428463825856 0.39352125162258744 0.58594749122858036 1.0217337106587363 0.61529834894463409 0.44165979465469707 0.14779896929394487 0.74544450687244523 0.34441631310619414 0.63463310385122895 0.54682576737832256 0.37520591542124748 0.72158903954550602 0.81367444177158155 0.52807081351056695 0.57383179012685992 0.92834634915925673 0.88328964146785327 0.36484908615239398 0.49441531533375382 0.30082511238288168 0.79381642222870119 1.0518765681190414 0.37338537507457659 1.2824593289988109 0.79741915501654148 0.49211289547383352 0.47080960555467744
leaf_count=36 25 24 65 21 17 40 20 20 12 32 28 26 30 23 33 26 30 38 45 45 21 14 25 55 72 41 41 41 24 30
internal_value=0 0.00173381 0.0386983 -0.00258196 0.00191295 0.0230453 -0.0255332 0.0448221 -0.00905712 0.00271407 0.00985405 -0.0023544 0.0237443 -0.0347363 0.00987302 0.00789675 -0.0164612 -0.00911471 -0.0660402 -0.0570678 -0.0146195 -0.0366535 -0.0484345 0.0737786 -0.0561323 -0.0291376 0.0800676 0.0795633 -0.0740373 -0.0676561
internal_weight=0 18.71 1.95609 16.7539 15.4771 5.28884 1.63703 3.65181 10.1883 7.84357 7.09812 5.95989 3.30012 2.65977 1.09679 1.78341 1.34175 1.03204 1.56298 1.27681 1.95587 1.35138 2.3447 1.13823 2.1969 1.14502 1.86841 1.34424 0.856962 0.771635
internal_count=1000 964 127 837 771 194 60 134 577 397 365 282 155 127 56 76 56 62 71 66 84 59 180 83 168 96 58 71 45 55
is_linear=0
shrinkage=0.1
Tree=69
num_leaves=31
num_cat=0
split_feature=0 1 5 8 4 6 11 6 7 6 10 9 1 15 30 25 10 3 1 9 0 6 1 27 1 14 10 5 0 4
split_gain=0.24832 0.333756 0.499008 0.367173 0.429791 0.41814 0.338484 0.427093 0.451913 0.377442 0.366718 0.328373 0.339148 0.302969 0.284618 0.279408 0.240662 0.236047 0.322703 0.235257 0.23207 0.214645 0.206869 0.190608 0.186218 0.178986 0.153201 0.143532 0.133507 0.0553383
threshold=-1.608339190483093 -2.1899367570877071 -1.4055383801460264 -0.82658091187477101 0.45788866281509405 -4.5184216499328604 4.6643924713134775 1.2706256508827212 0.25074130296707159 -0.35840791463851923 -0.43754623830318445 -1.7018492817878721 1.6095926165580752 -0.61889353394508351 -1.0374191403388975 0.48559746146202093 -2.9740501642227168 0.37997007369995123 -1.1795356869697569 1.1326313614845278 0.96046787500381481 -0.34531930088996882 -0.015287518501281737 0.27410019934177404 0.082951605319976821 0.98928418755531322 -1.0355508327484129 -1.0295194983482359 -0.29508113861083979 -3.2548031806945796
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=14 2 -2 4 5 -3 7 8 10 16 13 -8 -13 -5 -1 24 -10 -16 -19 21 22 -4 -12 -15 -6 -7 -9 -20 -27 -28
right_child=1 3 19 6 15 25 11 26 9 -11 20 12 -14 23 17 -17 -18 18 27 -21 -22 -23 -24 -25 -26 28 29 -29 -30 -31
leaf_value=0.050162043876156498 -0.054420673391815957 -0.099031646321335515 -0.014742990320066299 -0.0010409424492330954 0.069638604403575194 0.10113676568367214 0.087216282215667934 -0.0037270339400016514 0.01954410173743162 -0.038252275163724286 -0.0083174266439827709 0.038069776954483631 -0.059804400453325871 -0.091549755404258457 -0.091115464430163162 -0.070814346345120432 0.099184240868164963 -0.10208655814791034 -0.028162212985966856 0.10309341720771546 -0.044409975383778055 0.073300635957551305 0.07231026094535474 -0.031869867313418773 -0.017210641570404875 -0.045421640000791644 -0.043814339186233779 0.061392220693613192 0.07335006842601359 -0.096969954493920096
leaf_weight=0.35057545034214888 0.46192164695821736 0.14274340867996227 0.52421714388765395 1.0258196895010785 0.43618260184302904 1.08000312361401 0.60950772860087488 0.28672989457845699 0.8406702894717456 0.59680852852761734 0.6129782455973326 0.91252093669027157 0.57848069258034229 1.2916232594288892 0.74966459930874663 0.5029972344636916 0.69159725401550542 0.31330570776481181 0.33611133950762451 0.7919134615221991 0.54532955028116692 0.58692552824505273 0.66176739358341752 0.91376360552385438 0.56885996740311384 0.16737446375191212 0.25023010978475202 0.38279122160747625 0.21778784983325761 0.90124449424911257
leaf_count=18 19 4 15 34 21 154 24 11 38 34 29 28 30 62 63 23 31 30 24 37 21 26 24 39 25 9 15 17 32 63
internal_value=0 0.00409865 0.0388149 -0.0018358 0.0283796 0.0643471 -0.0106192 -0.0193919 -0.00942914 0.0292129 -0.0257165 0.0253761 9.64302e-05 -0.0459386 -0.0322024 -0.00996971 0.0554901 -0.0484073 -0.0173895 0.0614457 0.0101842 0.0317632 0.0335394 -0.0668224 0.0204814 0.0802642 -0.069132 0.0195225 0.0217372 -0.0854186
internal_weight=0 16.2 2.36498 13.835 3.11595 1.60791 10.7191 8.61856 7.18036 2.12908 5.05128 2.10051 1.491 3.23121 2.13245 1.50804 1.53227 1.78187 1.03221 1.90306 1.82008 1.11114 1.27475 2.20539 1.00504 1.46517 1.4382 0.718903 0.385162 1.15147
internal_count=1000 848 97 751 268 199 483 401 312 103 209 82 58 135 152 69 69 134 71 78 74 41 53 101 46 195 89 41 41 78
is_linear=0
shrinkage=0.1
Tree=70
num_leaves=31
num_cat=0
split_feature=30 30 0 8 1 6 2 10 0 5 28 0 1 4 8 6 11 8 5 7 3 4 5 5 9 3 7 8 3 15
split_gain=0.253314 0.327849 0.302491 0.380414 0.337863 0.293263 0.412071 0.327258 0.316289 0.315897 0.289594 0.289855 0.273685 0.258516 0.335261 0.258048 0.246646 0.290591 0.385779 0.279927 0.262014 0.258607 0.403196 0.215675 0.206188 0.186496 0.153619 0.13441 0.111454 0.10795
threshold=-1.2419461011886594 -1.1328976750373838 -1.608339190483093 -2.4568742513656612 3.704098224639893 2.090929508209229 3.1123524904251103 -4.9047172069549552 2.5667488574981694 -1.5503508448600767 -0.69158238172531117 1.0414988398551943 1.4367862939834597 1.4679937958717348 0.84256586432456981 -2.0005699396133418 3.1077305078506474 1.5303894281387331 -0.36188814043998713 0.30830553174018865 1.3480126261711123 -0.53900733590126026 0.38618937134742742 0.33765733242034918 0.4129705429077149 0.74455237388610851 -4.6473515033721915 0.3099705576896668 -0.33608692884445185 0.30655063688755041
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=13 -2 12 26 5 7 -7 15 10 -10 11 25 23 14 28 -5 17 19 -19 20 21 24 -23 -3 -12 -9 -4 -27 -1 -18
right_child=1 2 3 4 -6 6 -8 8 9 -11 16 -13 -14 -15 -16 -17 29 18 -20 -21 -22 22 -24 -25 -26 27 -28 -29 -30 -31
leaf_value=-0.010346755777196967 0.08025807050280806 -0.091830424075611178 -0.00493642439051197 0.015554099065458172 -0.077728941168493326 -0.07438433900884299 0.044147786874668728 0.010495075322641392 -0.087649824115031139 0.0095278898016429581 -0.00039819733161822304 0.036530424567179198 0.034523941504493387 -0.09137246617190381 -0.068602544276947397 -0.07961120572264907 0.095219623295583422 -0.078678886853326732 0.03219717426108476 0.092206506401238106 0.079418136450627591 -0.078390944770270532 0.04709853224823396 -0.012692141556378007 0.090009006707155448 -0.021530453870541734 0.10156185965518612 -0.10334388112096385 0.084455597863985071 0.039533629330300898
leaf_weight=0.20238087303005159 0.53990111046005473 0.81127634015866046 0.17856446793302894 0.49451373121701181 0.50933692674152631 0.95770004868973113 0.42276065051555634 0.51427653373684401 0.53404159191995859 0.89532340294681378 0.52609401242807219 0.91629417613148678 0.41074119112454344 0.58090385515242815 0.47076866449788213 0.67230960074812063 0.82427113305311628 0.56033591972663999 0.71325558610260464 0.7755380839225835 0.65185121563263226 0.63373166508971779 0.42960048699751496 0.598365438869223 0.48466371593531221 0.36547767813317478 0.56088588491548108 0.44568724907003343 0.32024015177740028 0.60264078481122851
leaf_count=11 24 88 8 24 23 64 19 32 28 35 25 37 20 23 20 32 30 35 26 72 39 30 18 31 23 17 90 23 34 19
internal_value=0 0.00365615 0.000986126 0.00608586 0.00209637 0.00536987 -0.0380843 0.0108035 0.0167224 -0.0267798 0.0240865 -0.0067171 -0.0373075 -0.0383807 -0.00739271 -0.039279 0.0352206 0.0243192 -0.0165844 0.0391971 0.0241158 0.00673517 -0.0276915 -0.0582378 0.0429525 -0.0366147 0.0758444 -0.0664821 0.0477441 0.0717012
internal_weight=0 16.0294 15.4895 13.6692 12.9297 12.4204 1.38046 11.0399 9.87308 1.42936 8.44372 2.24174 1.82038 1.57429 0.99339 1.16682 6.20198 4.77507 1.27359 3.50148 2.72594 2.07409 1.06333 1.40964 1.01076 1.32544 0.73945 0.811165 0.522621 1.42691
internal_count=1000 912 888 749 651 628 83 545 489 63 426 109 139 88 65 56 317 268 61 207 135 96 48 119 48 72 98 40 45 49
is_linear=0
shrinkage=0.1
Tree=71
num_leaves=31
num_cat=0
split_feature=18 18 8 1 11 7 5 6 20 30 8 6 10 18 23 4 1 20 0 27 26 11 21 11 4 9 25 8 19 6
split_gain=0.234848 0.290091 0.268183 0.277724 0.253848 0.376568 0.343351 0.403537 0.319758 0.333075 0.319186 0.285981 0.300287 0.283602 0.267255 0.256651 0.405532 0.295054 0.225448 0.222974 0.203747 0.19815 0.184038 0.203891 0.167502 0.242386 0.164904 0.155452 0.121365 0.12058
threshold=-1.5012420415878294 -0.79747301340103138 -3.1696652173995967 -3.7169742584228511 -0.97970148921012867 0.25074130296707159 -0.67728805541992176 -1.5258474349975584 0.36504785716533666 0.22748384624719623 -1.1171276569366453 0.32099404931068426 -0.93400558829307545 0.93944856524467479 1.0135984420776369 1.3520237207412722 1.8695499300956728 -0.32131305336952204 0.77826052904129039 -0.78124374151229847 -0.39324674010276789 2.3344395160675053 0.10897739604115488 0.95515820384025585 -0.92380249500274647 -1.5438122153282163 -0.76631933450698841 0.55232834815979015 0.82704880833625805 1.1451571583747866
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=19 21 -3 -4 5 10 7 -6 13 27 -5 22 -13 15 28 16 17 -9 -14 -1 -7 24 23 -8 25 -2 -22 -10 -12 -23
right_child=1 2 3 4 6 20 11 8 9 -11 14 12 18 -15 -16 -17 -18 -19 -20 -21 26 29 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.028259086859171247 -0.097818487654398656 0.07633251226485277 0.067504702894146659 0.018985243831281876 -0.10240885714058581 -0.027092046439376873 -0.019203718256808409 0.061466622046986857 0.029980355158744733 0.076299431045597332 -0.097744827525571273 -0.055074595751849698 -0.023422812034188559 0.04323935462263561 -0.0031916323128535285 -0.086156670378302894 -0.08524940772269711 -0.05130662705161107 0.088629518060621804 0.06388536658753928 -0.021062589518493324 0.038383373771455252 0.085234312372324622 0.071022349388357262 -0.08371234909963228 0.038651001084098462 0.088896317620540011 -0.056687702078270168 -0.028502991821523483 -0.039774858270065501
leaf_weight=0.35575686884112645 0.21231269184499979 0.46922074392205182 0.57129939121659745 0.60724661732092577 0.43732785992323986 0.47610059357248247 0.51647452358156376 0.66756395157426285 0.42132743215188362 0.85860952350776643 1.2730346486205211 0.63207826227881014 0.32163460855372228 0.5347660679835825 0.55169327044859517 0.73888023430481542 0.53375591011717904 0.35557485464960803 0.40648553951177746 1.0030287727713609 0.18909208872355521 0.44438431109301735 0.99630398268345732 0.4862637355690822 0.93893238587770711 0.33629875665064912 0.48931165027897783 0.40675154747441411 0.31596508971415449 0.35513806925155222
leaf_count=16 15 63 31 40 34 39 23 35 23 37 93 23 23 33 23 29 28 26 24 52 14 25 41 20 61 28 42 18 22 19
internal_value=0 -0.00359259 0.00208166 -0.000642868 -0.00382983 -0.02487 0.00604631 -0.0106872 -0.00180734 0.0326588 -0.0450048 0.0307267 -0.00464522 -0.0223453 -0.0631568 -0.0376223 -0.0145885 0.0222742 0.0391323 0.0397601 0.0230547 -0.0364828 0.0547946 0.0245501 -0.0580622 -0.0141627 0.0582474 -0.0125909 -0.0839764 0.00366644
internal_weight=0 15.5438 13.2568 12.7875 12.2162 3.90244 8.3138 4.95456 4.51723 1.68669 2.74794 3.35924 1.3602 2.83054 2.14069 2.29577 1.55689 1.02314 0.72812 1.35879 1.1545 2.28707 1.99904 1.00274 1.48754 0.548611 0.678404 0.828079 1.589 0.799522
internal_count=1000 932 784 721 690 273 417 263 229 78 178 154 70 151 138 118 89 61 47 68 95 148 84 43 104 43 56 41 115 44
is_linear=0
shrinkage=0.1
Tree=72
num_leaves=31
num_cat=0
split_feature=8 3 4 9 9 9 8 11 16 2 16 5 27 19 28 1 0 18 19 4 6 10 14 13 22 4 0 15 31 6
split_gain=0.245317 0.369463 0.318269 0.299895 0.343278 0.296904 0.379531 0.341662 0.28955 0.371068 0.345412 0.416986 0.396279 0.398461 0.278267 0.2746 0.346352 0.325419 0.317878 0.265887 0.240708 0.216558 0.199473 0.192353 0.158082 0.152326 0.136039 0.133078 0.132936 0.106752
threshold=0.93145474791526806 -2.818352103233337 -0.16142201423645017 -1.4426879882812498 0.26169002056121832 -0.4592334628105163 -0.82658091187477101 1.9994604587554934 0.80493414402008068 3.0034747123718266 0.54110366106033336 -0.93297564983367909 -1.0113512277603147 0.55610954761505138 0.87583532929420482 0.47342592477798467 -0.62135785818099964 0.091590221971273436 0.13681121915578845 1.2404614686965945 -3.4843884706497188 -2.7094286680221553 -0.69454568624496449 -0.012470239773392676 -2.417659997940063 -2.2204345464706416 -1.4239488244056699 0.24113584309816363 0.59073963761329662 2.0313966274261479
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 -1 5 23 27 6 26 21 9 10 11 14 -13 22 25 16 17 -6 -17 -18 -7 -8 -14 -4 -22 -2 -3 -5 -10 -26
right_child=8 2 3 4 15 20 7 -9 28 -11 -12 12 13 -15 -16 18 19 -19 -20 -21 24 -23 -24 -25 29 -27 -28 -29 -30 -31
leaf_value=-0.072728411738747431 -0.017459384912261563 -0.0032571693967986867 0.0037814284077032191 -0.092953408579440178 0.025698953470520119 -0.029131488562386766 0.012736682737632952 0.056023242789790963 -0.092304403358462461 0.066490512513505171 0.057140956076721795 0.073726664708306966 0.0082127022815200231 0.042833890962467036 0.020057679700930954 -0.088788416337938714 7.0745631597230816e-05 -0.089231770825314957 0.012471744346584717 0.09227336389968685 -0.10249907609701958 -0.079217864025685772 -0.079558343055587788 0.10352285029309699 0.09420322968219727 -0.10226809864095432 0.10172467812783775 -0.017097530635597071 -0.018021559120444848 0.018304237244997343
leaf_weight=0.51606710953638035 0.26502269832417291 0.165797419846058 0.37192082009278238 0.60751959308981884 0.60837230423931055 0.2384769013733602 0.40037919185124327 0.46478533628396679 0.63700097368564434 0.6210657360497861 0.56988422130234528 0.56268885638564858 0.36355072725564241 0.63440043246373523 0.30999235855415452 0.58674525795504429 0.509278135141358 0.41401658090762794 0.6573190416675061 0.81051761889830232 0.046873427927493494 0.71076377923600365 0.89976458507589929 0.4027113914489745 1.3597499309107663 1.0542644535889849 0.48307966481661446 0.37343796831555665 0.38745128898881376 0.21455271472223092
leaf_count=39 16 7 20 27 30 15 17 21 50 20 19 18 16 21 18 34 26 30 25 36 2 43 52 26 138 80 78 26 23 27
internal_value=0 0.00965849 0.014169 -0.00189852 -0.0116565 0.0351828 0.0105331 -0.0159696 -0.0155569 -0.00611808 -0.015796 -0.0259595 2.47788e-05 -0.0218285 -0.0651988 0.00268192 0.0228489 -0.0208423 -0.0352862 0.0566945 0.0646725 -0.0460838 -0.0543 0.0556344 0.0784712 -0.0852314 0.0749003 -0.064076 -0.0642104 0.0838594
internal_weight=0 9.94236 9.4263 5.34184 4.56721 4.08446 2.22481 1.57593 6.30509 5.28063 4.65957 4.08968 2.4604 1.89772 1.62928 3.58625 2.34218 1.02239 1.24406 1.3198 1.85965 1.11114 1.26332 0.774632 1.62118 1.31929 0.648877 0.980958 1.02445 1.5743
internal_count=1000 667 628 280 234 348 166 81 333 260 240 221 107 89 114 181 122 60 59 62 182 60 68 46 167 96 85 53 73 165
is_linear=0
shrinkage=0.1
Tree=73
num_leaves=31
num_cat=0
split_feature=8 7 6 6 11 5 7 1 6 19 31 2 11 2 3 3 8 6 1 22 27 31 29 19 4 19 29 20 12 30
split_gain=0.210748 0.336326 0.294999 0.289483 0.262281 0.304244 0.323836 0.31563 0.355578 0.447358 0.304259 0.237437 0.330561 0.325018 0.330614 0.212241 0.20389 0.354638 0.314954 0.185969 0.181113 0.169316 0.165714 0.157661 0.147724 0.135674 0.160501 0.128848 0.101686 0.0897148
threshold=0.93145474791526806 -0.86519092321395863 1.9243513345718386 1.1906984448432925 -5.2955746650695792 -0.93297564983367909 -1.4436311721801756 -2.1634254455566402 -2.7600373029708858 0.82704880833625805 -0.38416753709316248 2.9853472709655766 1.9153140187263491 0.73140716552734386 -0.9574308693408965 2.0606999397277836 -0.5797320008277892 -0.80244350433349598 3.3072034120559697 0.60016271471977245 -0.6061393916606902 0.42323160171508795 -0.27872712910175318 0.32967743277549749 1.2241573929786684 -0.56584131717681874 0.12018478289246561 0.44136221706867224 0.036220900714397437 -0.86477023363113392
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 2 3 11 -2 6 22 -7 -9 19 27 12 13 14 -1 16 18 -18 -3 24 -4 -5 -6 -15 -10 -19 -27 -8 -22 -16
right_child=4 15 20 21 5 7 10 8 9 -11 -12 -13 -14 23 29 -17 17 25 -20 -21 28 -23 -24 -25 -26 26 -28 -29 -30 -31
leaf_value=-0.053428922511919066 -0.1026839822264884 0.079205967756128245 0.017521293487888733 0.098196897283693421 0.050816545955621656 0.073880539672267528 -0.052388059219618859 0.070768933432948627 -0.10196184241592361 0.058835215946172463 -0.10185985725730225 -0.066189403982579054 0.045564444730139236 -0.0897106278542944 0.10410664306965584 0.081428687458509466 0.07532110376167496 0.021506199031454686 -0.067928392299921522 0.00011221743058167881 -0.098737482653857234 0.0052790244783442256 -0.032016940458483613 -0.0084245911963247381 -0.020073047022025257 0.0010929440031314445 -0.085142411323606504 0.031136774471121134 -0.031641789173801026 0.030379318425323459
leaf_weight=0.39746057987213068 0.32110814470797766 0.89769168861676074 0.30905728251673281 0.48141279281117011 0.5000255664344877 0.63992158602923188 0.41524451063014567 0.49637250136584099 0.68371873802971117 0.49136033188551675 0.88573356741108 0.54970001988112915 1.2445676713250575 0.64740120910573729 0.28545074595604025 0.73362145281862456 0.45926253602374334 0.39678780804388214 0.17362378153484315 0.47864453587681044 0.54326728649903067 0.33090946555603296 0.46716102561913442 0.37789119558874518 0.32501104287803173 0.37258976383600284 0.51297226455062628 0.33264574338681996 0.38662917981855571 0.39128934813197858
leaf_count=29 23 126 18 28 19 24 29 17 49 16 83 29 53 37 12 93 32 21 8 22 38 22 24 19 13 23 25 14 24 30
internal_value=0 0.00915485 -0.00538422 0.00604537 -0.0147435 -0.00980315 -0.0350524 0.011278 -0.00490744 -0.0238911 -0.0622034 -0.00528284 0.00472905 -0.0194779 0.0189616 0.0335265 0.0210335 -8.20602e-05 0.0553605 -0.0512201 -0.0487988 0.0603457 0.0108071 -0.0597511 -0.0755774 -0.0270871 -0.0488599 -0.015238 -0.0708407 0.0614777
internal_weight=0 9.49159 5.94504 4.70608 6.03695 5.71584 2.60081 3.11503 2.47511 1.97873 1.63362 3.89376 3.34406 2.09949 1.0742 3.54655 2.81293 1.74161 1.07132 1.48737 1.23895 0.812322 0.967187 1.02529 1.00873 1.28235 0.885562 0.74789 0.929896 0.67674
internal_count=1000 667 339 259 333 310 169 141 117 100 126 209 180 127 71 328 235 101 134 84 80 50 43 56 62 69 48 43 62 42
is_linear=0
shrinkage=0.1
Tree=74
num_leaves=31
num_cat=0
split_feature=4 9 9 11 5 6 9 5 7 26 21 7 5 4 3 13 14 8 29 8 17 3 27 6 21 17 8 18 30 10
split_gain=0.191709 0.553927 0.358129 0.355701 0.314845 0.431754 0.464525 0.389698 0.342931 0.293239 0.28712 0.284229 0.310848 0.288382 0.280069 0.272742 0.259218 0.221476 0.257868 0.191797 0.189551 0.162948 0.200284 0.134544 0.130687 0.121455 0.11955 0.113746 0.10468 0.0706225
threshold=-1.0175390839576719 -2.2834039926528926 -1.9159966707229612 0.91672056913375866 1.4445725679397585 -1.5970541238784788 1.2183623313903811 -2.3675128221511836 -2.577299833297729 -0.82819029688835133 0.60416230559349071 -1.3801147341728208 -0.95051786303520192 1.6306554079055788 -0.18057116866111753 -1.2598152160644529 -0.57414984703063954 0.41712841391563421 -0.031061415560543534 -0.52487343549728382 -1.1937925815582273 0.30778414011001592 -0.42488506436347956 -1.6460761427879331 0.12501231953501704 0.68529933691024791 1.4684590697288515 -0.088103394955396638 0.28474009037017828 -0.66635462641715992
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 10 3 -2 5 20 7 27 -6 16 -1 13 19 -8 17 -9 -3 29 -19 -13 -4 -17 -23 -10 -5 -18 -16 -7 -24 -11
right_child=2 9 4 24 8 6 11 15 23 14 -12 12 -14 -15 26 21 25 18 -20 -21 -22 22 28 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.079920440841156237 -0.044005240349456583 -0.098988486653481256 0.01128210451979679 0.10432895392476733 -0.050710331316338508 0.074294759620133891 0.08961431611971532 0.032493746310536774 0.10447078742912018 0.078586855976468223 0.033754936262749988 0.0076856998984036603 0.037003034062401317 0.0013011273866689777 0.093720009874784779 -0.089874404132185962 0.065237222551711188 0.029998161796337453 -0.082834654957245582 -0.10323024864879571 -0.10133430438644631 0.022523075237865023 -0.033651219359389521 0.019650697103623847 0.028475661142486525 -0.038594224952169714 0.040462190454493979 -0.0084471635851992109 -0.10124795160887801 0.013858607193472403
leaf_weight=0.78355880087474339 0.35962631250731536 0.28517307923175395 0.17983922432176769 0.54975680029019713 0.33622099622152746 0.32289575692266226 0.77005102182738505 0.32908796262927165 0.46265014773234725 0.4535096368053928 0.31013904977589846 0.24338274903129786 0.56695347989443667 0.71130633703432977 1.0618585312913649 1.2578331718687028 0.27151011588284735 0.44372502574697137 0.37265291973017156 0.43374791066162288 0.88473802461521411 0.37271813699044287 0.50506746210157849 0.3138892459683118 0.38704353768844157 0.19255106907803565 0.69890586449764658 0.3422452891245461 0.41926750825950876 0.26827148278243818
leaf_count=60 22 13 15 28 19 17 39 17 21 39 20 13 31 37 128 89 27 20 23 30 83 23 31 27 22 19 27 23 17 20
internal_value=0 0.0154605 -0.00840371 0.0405356 -0.0159105 -0.0234258 -0.0134352 -0.0372788 0.0336566 0.0325208 -0.0476856 0.0176143 -0.0176245 0.0472088 0.0453472 -0.0531922 -0.0239547 0.0141729 -0.0215068 -0.0633635 -0.08231 -0.0642292 -0.0393595 0.0701852 0.0729898 0.0221549 0.0725802 0.0317203 -0.0643123 0.0545287
internal_weight=0 5.14186 9.74832 1.29643 8.45189 7.33913 6.27456 3.54912 1.11276 4.04816 1.0937 2.72544 1.24408 1.48136 3.29892 2.88397 0.749234 1.53816 0.816378 0.677131 1.06458 2.55489 1.29705 0.776539 0.9368 0.464061 1.76076 0.665141 0.924335 0.721781
internal_count=1000 396 604 72 532 465 367 217 67 316 80 150 74 76 257 177 59 102 43 43 98 160 71 48 50 46 155 40 48 59
is_linear=0
shrinkage=0.1
Tree=75
num_leaves=31
num_cat=0
split_feature=18 8 16 4 4 9 1 0 9 20 3 29 0 5 9 7 10 7 11 3 10 6 6 22 13 21 8 10 19 4
split_gain=0.190237 0.244207 0.352285 0.408143 0.272033 0.32401 0.293207 0.297971 0.276944 0.235407 0.235311 0.294519 0.253104 0.234106 0.229532 0.221109 0.210715 0.209289 0.192632 0.179722 0.177611 0.348539 0.169689 0.168323 0.161002 0.147489 0.0994916 0.0978174 0.0860817 0.0559625
threshold=-1.5012420415878294 1.7825397253036501 -0.63479617238044728 0.73517119884490978 -0.22178694605827329 -1.8577739000320432 0.77400022745132457 -0.641687512397766 0.46225464344024664 0.075913507491350188 -0.75348734855651844 -0.031061415560543534 1.3665547370910647 -0.15890216827392575 -1.6104571819305418 -1.0387114286422727 -3.102136373519897 -0.73147648572921742 -1.4468333125114439 0.46754512190818792 -1.1965752840042112 0.42232704162597662 -2.9369398355484004 -0.71317166090011586 0.452407255768776 0.24724747240543368 2.9123635292053227 4.4430315494537362 0.50487545132637035 -3.7941824197769161
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=19 4 24 29 10 25 7 9 -9 13 11 17 14 -7 18 28 -13 -2 -12 -1 -8 -22 -16 -24 -3 -6 -5 -25 -10 -4
right_child=1 2 3 26 5 6 20 8 15 -11 12 16 -14 -15 22 -17 -18 -19 -20 -21 21 -23 23 27 -26 -27 -28 -29 -30 -31
leaf_value=0.0020423909408283339 -0.003462012861259617 -0.021680181869639902 -0.046347147362766086 -0.03079880997934864 0.10040648273313467 -0.063127794241805826 -0.090900571927672424 -0.035676903090370304 0.10158507957222383 -0.097429394748412002 -0.086208565640998769 0.020578450150715782 0.089389529911552662 0.04302828551444865 -0.04806413487666996 -0.0044686998532900485 -0.079578233423459543 0.097102246241114576 0.034724881341698211 0.08049088678134815 0.0483233617066725 -0.074036365934129864 -0.015406279636917267 0.089161348051564615 0.067112300660162313 0.0076357086701759171 0.044666846772222261 0.011975839170351391 0.039671349854547774 -0.10176140830229341
leaf_weight=0.63057819788809932 0.53710057726129889 0.39568972191773388 0.22638032212853432 0.36084366962313641 0.3389381180750205 0.404789874330163 0.72959417453967035 0.62315947259776372 0.51178118819370866 0.46603781776502728 0.30790627747774124 0.31857902160845697 0.75289709144271899 0.42675487161613979 0.18348640151088869 0.58209666167385854 0.61663803295232156 0.33666480786632746 0.2301819720305488 0.54394440015312284 0.38509961590170838 0.58861902169883251 0.27000790293095445 0.80672931176378015 0.42200156394392252 0.34663237212225795 0.33865170046919957 0.20614475308684632 0.400136694079265 0.93475883163045981
leaf_count=37 30 28 15 26 24 31 53 25 24 45 25 15 93 21 12 34 40 25 23 31 15 45 27 84 14 17 15 27 20 79
internal_value=0 -0.00364434 -0.0305634 -0.0546058 0.00330822 -0.0110584 -0.0197062 -0.00280168 0.0203241 -0.0405344 0.0215677 -0.00645868 0.0399546 -0.00864771 0.0213862 0.0436824 -0.0454601 0.0352857 -0.034476 0.0383734 -0.0535958 -0.0256439 0.041885 0.0547502 0.0241447 0.0535005 0.00573692 0.0734522 0.0744182 -0.0909576
internal_weight=0 13.0483 2.67833 1.86063 10.37 5.80364 5.11807 3.41476 2.11717 1.29758 4.56634 1.80898 2.75735 0.831545 2.00446 1.49401 0.935217 0.873765 0.538088 1.17452 1.70331 0.973719 1.46637 1.28288 0.817691 0.68557 0.699495 1.01287 0.911918 1.16114
internal_count=1000 932 177 135 755 354 313 200 103 97 401 110 291 52 198 78 55 55 48 68 113 60 150 138 42 41 41 111 44 94
is_linear=0
shrinkage=0.1
Tree=76
num_leaves=31
num_cat=0
split_feature=30 30 0 8 1 20 1 27 25 24 0 22 11 7 9 1 27 18 23 5 2 27 8 22 4 21 9 8 16 25
split_gain=0.183895 0.216987 0.227624 0.281625 0.235955 0.227095 0.228729 0.267976 0.202567 0.259152 0.212273 0.199199 0.192341 0.249608 0.268859 0.184407 0.16835 0.23765 0.282447 0.167879 0.148102 0.137347 0.286837 0.265392 0.134056 0.132992 0.106638 0.146575 0.0934855 0.125632
threshold=-1.2419461011886594 -1.1328976750373838 -1.608339190483093 -2.5952844619750972 -3.7169742584228511 0.35997511446475988 0.77400022745132457 0.60264223814010631 -0.64878144860267628 0.11370582506060602 2.3491891622543339 0.57570406794548046 0.49007196724414831 0.7070046365261079 0.19106149673461917 1.4367862939834597 -0.59541326761245716 -1.3968747258186338 1.4596264958381655 0.33765733242034918 1.6277096867561343 -0.44899147748947138 1.4811351895332339 0.18504730612039569 0.29749429225921636 -1.0671396255493162 -1.1705294847488401 0.22274678945541385 0.086818952113389983 0.49037462472915655
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=11 -2 15 -4 -5 6 8 25 9 -6 16 24 13 14 -7 19 20 -18 28 -3 -10 -14 23 -23 -1 -8 27 -27 -19 -30
right_child=1 2 3 4 5 12 7 -9 10 -11 -12 -13 21 -15 -16 -17 17 18 -20 -21 -22 22 -24 -25 -26 26 -28 -29 29 -31
leaf_value=-0.023719757898985783 0.075613206547818113 -0.090735055054332187 0.080484749583397061 0.06884730133697646 -0.0050917342451794346 -0.072478396662500472 0.0065046510767515393 0.030472370239603627 0.0012707192968031014 0.088174797508560207 -0.06790208603856239 0.022538359791356623 0.074348538468407568 0.059993725834630501 0.033464581258186063 0.030190380145644364 0.086244344787908961 -0.085469140280738043 0.051153821438125524 -0.012019918011607383 0.097088205110152745 0.094474948235161643 -0.058632174476759592 -0.0088940010726111449 -0.10391866623567372 0.012816206891185185 -0.092380239319667143 -0.090073460317697085 -0.074046429974926192 0.024040143595830721
leaf_weight=0.40819197849487021 0.40331382065778498 0.61440376134123542 0.48541309061692572 0.50916617759502969 0.68354125437326174 0.66537307610269658 0.27158720145234849 0.43645751709118474 0.36955200997180759 0.52809216966852546 0.56902564293704916 0.38182522065471858 0.71439419069793109 0.38415561214787874 0.37428973673377186 0.32043782621622074 0.19469979382119573 0.54287565237611735 0.31670479336753499 0.48468517849687487 0.28627678542397916 0.5428667342639526 0.34433612134307612 0.45785400358727202 0.42588452203199267 0.29216708778403699 0.84302707883762196 0.26317726518027484 0.26779370935400948 0.25485127221327275
leaf_count=32 24 88 89 27 31 49 21 36 29 24 31 32 31 35 32 20 14 46 9 31 17 36 22 23 24 22 64 21 22 18
internal_value=0 0.00346472 0.00104332 0.00608034 0.00250875 -0.00100859 -0.0126103 -0.0392958 0.00139544 0.0355586 -0.0133785 -0.0372841 0.0193747 -0.00888652 -0.0343378 -0.0365613 0.00051707 -0.0171912 -0.0317611 -0.0560226 0.0430961 0.0389134 0.0200928 0.0471811 -0.0646698 -0.0575303 -0.069967 -0.0359431 -0.0564059 -0.0262176
internal_weight=0 12.4205 12.0172 10.5977 10.1123 9.6031 6.11983 2.10642 4.01341 1.21163 2.80178 1.2159 3.48327 1.42382 1.03966 1.41953 2.23275 1.57693 1.38223 1.09909 0.655829 2.05945 1.34506 1.00072 0.834077 1.66996 1.39837 0.555344 1.06552 0.522645
internal_count=1000 912 888 749 660 633 405 164 241 55 186 88 228 116 81 139 155 109 95 119 46 112 81 59 56 128 107 43 86 40
is_linear=0
shrinkage=0.1
Tree=77
num_leaves=31
num_cat=0
split_feature=6 0 15 8 31 12 4 11 5 6 8 7 1 9 11 19 19 12 30 9 3 9 3 30 10 3 29 0 10 0
split_gain=0.17698 0.222072 0.304319 0.245298 0.20852 0.270283 0.352428 0.344227 0.339372 0.447955 0.319571 0.284103 0.277435 0.318772 0.252035 0.225775 0.220076 0.211866 0.200818 0.18706 0.183419 0.179993 0.167523 0.159509 0.159292 0.151937 0.150549 0.181292 0.157865 0.106412
threshold=2.2315721511840825 0.46007883548736578 -0.8999110460281371 0.11012277007102968 1.3691098093986513 0.079420387744903578 -1.1762363910675047 -1.0592629313468931 -0.89550951123237599 -0.28584939241409296 0.063155800104141249 -1.6642212867736814 -0.6302025020122527 -1.1534116268157957 3.7105647325515752 0.32967743277549749 1.2757573723793032 -0.24935476481914517 -1.0102906227111814 0.023247927427291874 0.78698328137397777 -2.3239196538925166 -0.40226960182189936 -0.15728227049112317 -0.75833630561828602 0.66349661350250255 0.25818875432014471 -0.99946826696395863 -3.4719688892364498 0.37160789966583258
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=4 -2 -3 -4 5 6 16 25 9 -9 11 22 -10 -14 19 23 18 -6 -1 29 -21 -20 -7 -11 -15 -8 27 -13 -29 -12
right_child=1 2 3 -5 17 10 7 8 12 15 14 26 13 24 -16 -17 -18 -19 21 20 -22 -23 -24 -25 -26 -27 -28 28 -30 -31
leaf_value=-0.013480336983546122 -0.089643141118086411 -0.096995426452445174 0.09113028363429998 -0.047799115592242918 -0.017142888594251116 -0.08910647520173802 -0.091400664157703038 -0.10248240692425184 0.099907740401044276 0.028511067930977366 -0.031457604816098821 -0.076223588589924535 0.073947968988848214 -0.077604404826739054 0.021841760482528637 0.065100629792320952 -0.039117062342112638 0.065897234727394199 -0.0090875768186425201 0.032951292062908706 -0.067775211363671581 0.090668078193832324 0.0048653776275446236 -0.071234089688178404 0.018731774185251182 0.00098959822680369449 0.095737766044898864 -0.034438593820145268 0.092019097777147771 -0.097729982769307708
leaf_weight=0.30750758785870846 0.47659688233397968 0.23879119101911783 0.33588338561821729 0.20444330945611 0.44986195507226512 0.31690863380208611 0.56408035464118889 0.49366047768853372 0.5420702521223576 0.29220304940826924 0.32108064665226332 0.15713675646111369 0.5048594246036372 0.35262206813786168 0.47738054988440115 0.46535939781460911 0.25885893328813825 0.96916062140371539 0.21287237695651118 0.35673627560026933 0.36652785312617198 1.2033624756149919 0.47262000699993223 0.35523150209337717 0.33441640716046694 0.26005711365723982 0.39200367993908003 0.14974550972692668 0.28969832486473002 0.98727780277840793
leaf_count=21 55 11 47 13 34 25 56 42 26 18 28 11 27 22 30 26 23 50 22 27 28 96 34 24 20 20 47 13 41 63
internal_value=0 -0.0358747 -0.00298378 0.0385636 0.00360579 -0.00128543 0.0121553 -0.00436549 0.0099148 -0.0232001 -0.0205577 0.0118742 0.0405945 0.013619 -0.0435419 0.0119713 0.0468581 0.0395717 0.0597692 -0.0589055 -0.0180938 0.0756739 -0.032854 -0.0262167 -0.0307127 -0.0622469 0.0475961 0.015963 0.0489272 -0.0814663
internal_weight=0 1.25571 0.779118 0.540327 11.8533 10.4343 6.14716 4.16456 3.34042 1.60645 4.28712 1.77811 1.73397 1.1919 2.509 1.11279 1.9826 1.41902 1.72374 2.03162 0.723264 1.41623 0.789529 0.647435 0.687038 0.824137 0.988584 0.596581 0.439444 1.30836
internal_count=1000 126 71 60 874 790 443 281 205 110 347 171 95 69 176 68 162 84 139 146 55 118 59 42 42 76 112 65 54 91
is_linear=0
shrinkage=0.1
Tree=78
num_leaves=31
num_cat=0
split_feature=5 7 8 22 11 1 3 8 9 0 4 9 4 9 3 9 6 22 21 28 27 4 17 25 27 7 11 19 6 8
split_gain=0.166991 0.16632 0.261416 0.245597 0.244142 0.254971 0.335809 0.271978 0.294965 0.310817 0.493154 0.25093 0.408098 0.315375 0.237556 0.216098 0.213801 0.213282 0.179998 0.161089 0.155238 0.138349 0.123914 0.120457 0.111762 0.104921 0.0851055 0.0791849 0.0773167 0.0619669
threshold=-4.0636310577392569 0.72382399439811718 2.3894670009613042 0.26117700338363653 0.32232411205768591 0.31563207507133489 1.3817856311798098 1.6661189794540407 -1.8577739000320432 0.8787231743335725 1.8425122499465945 -2.2834039926528926 -2.1853830814361568 -1.1903105974197385 2.016233086585999 -1.2095330357551572 -0.50517910718917836 0.26117700338363653 0.30136841535568243 -0.92239686846733082 0.29147735238075262 -0.83761519193649281 0.30233138799667364 -0.056806737557053559 0.959560126066208 -1.5995073318481443 2.8430706262588505 1.0498985648155215 1.1343235969543459 -1.2007194757461546
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=-1 4 3 15 11 6 19 8 -7 10 28 -2 13 -13 17 -3 -8 24 23 -6 21 -17 -18 -19 27 -11 -21 29 -10 -14
right_child=1 2 -4 -5 5 7 16 -9 9 25 -12 12 14 -15 -16 20 22 18 -20 26 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.072928406062737941 -0.080970041316288266 0.10132292371958736 -0.033229990985884937 0.091793064402409377 0.0086422390710978944 0.069730732521702324 -0.066472680884215457 -0.082281907955393113 0.080498382739907545 -0.096302633661899839 -0.074278870439188391 -0.038848608326135708 -0.042674529557299029 0.096934083882514982 0.041821752746617098 0.011968091770732568 0.057899349428582575 -0.09701964523707772 0.045351297086245412 0.042064395812356596 0.054439950500708302 -0.096812416878617594 -0.025955922956947354 -0.0018354917864821974 -0.0033797826310446527 -0.024986297406989391 0.10030470610178617 -0.023212911223856016 0.018915983356093763 -0.10177720557710479
leaf_weight=0.30775425606407303 0.61567860317881962 0.21166467876173567 0.5401489654323085 0.69783612951869145 0.41273661353625346 0.51930950256064612 0.41444627597229555 0.44557717512361694 0.56541694368934281 0.43206017534248531 0.41101707296911627 0.25458097987575457 0.22826726245693718 0.5213744499487788 0.4026768163894302 0.21675412525655713 0.42951271147467174 0.23017074028030038 0.41532231133896858 0.37063164124265335 0.22394602152053267 0.25383014755789102 0.29882514837663621 0.31478366930969048 0.24867506593000133 0.39479107240913425 0.77672023396007717 0.22603865177370597 0.31883622531313449 0.7960161919472778
leaf_count=36 50 37 36 76 15 36 28 37 29 28 28 20 19 62 28 28 30 18 30 24 34 17 20 28 23 28 35 16 26 78
internal_value=0 0.00165795 0.0269407 0.0472028 -0.00373966 0.00962391 0.0320507 -0.0100122 0.00217886 -0.0143519 0.0162255 -0.0219299 -0.0119379 0.0523855 -0.0293778 0.0128651 -0.00913307 -0.0410359 -0.00424197 0.0622183 -0.0140932 -0.0467074 0.0234949 -0.0420381 -0.0646065 -0.0622517 0.0814912 -0.0767838 0.0582935 -0.0886058
internal_weight=0 12.1876 2.14418 1.60403 10.0435 5.78988 2.70287 3.08701 2.64143 2.12212 1.29527 4.25358 3.63791 0.775955 2.86195 0.906195 1.14278 2.45927 0.960277 1.56009 0.69453 0.470584 0.728338 0.544954 1.499 0.826851 1.14735 1.25032 0.884253 1.02428
internal_count=1000 964 228 192 736 364 152 212 175 139 83 372 322 82 240 116 78 212 76 74 79 45 50 46 136 56 59 113 55 97
is_linear=0
shrinkage=0.1
Tree=79
num_leaves=31
num_cat=0
split_feature=6 11 4 2 8 3 27 0 2 7 5 21 26 2 26 30 30 0 15 3 4 19 2 28 21 0 0 30 28 3
split_gain=0.166938 0.22956 0.192162 0.201974 0.174509 0.245224 0.194658 0.181639 0.20024 0.211364 0.25136 0.208196 0.189249 0.168723 0.224416 0.18593 0.215812 0.162512 0.209144 0.243918 0.204642 0.153327 0.179056 0.145228 0.129 0.193494 0.123832 0.107733 0.0867282 0.0707613
threshold=1.7042843699455263 6.7042803764343271 -3.3705158233642574 3.0776867866516118 0.93145474791526806 -2.7369303703308101 -1.4307758212089536 1.4968117475509646 3.4516642093658452 0.11073687672615053 -1.9478189349174497 -0.85308596491813649 0.48127926886081701 3.0034747123718266 -1.0634013414382932 -1.2419461011886594 0.55339282751083385 2.7861731052398686 0.040439540520310409 -0.11346042156219481 -1.3023312687873838 0.59844911098480236 1.1631748676300051 -0.20270609110593793 0.10627708956599237 2.3296927213668828 0.68990981578826915 -0.13614627718925473 -0.11683378741145133 -1.429771304130554
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=4 2 -2 -4 5 -1 -7 8 9 10 -8 -12 21 14 -6 -16 17 18 19 -17 -21 22 27 -20 -9 -26 -18 -13 -14 -11
right_child=1 -3 3 -5 13 6 7 24 -10 29 11 12 28 -15 15 16 26 -19 23 20 -22 -23 -24 -25 25 -27 -28 -29 -30 -31
leaf_value=-0.079218639374103353 0.046332484879701707 0.051245674494542065 -0.071969479200536562 0.0055234554722132633 -0.096455628355673201 -0.049329169316061239 0.091465538288766857 0.099563777906623435 -0.056762604429350584 -0.017533103826682103 -0.060563566341201316 0.078973027455566558 0.0037737319154358775 0.052201191749809805 -0.084247987072627858 0.057114627277892641 -0.10159827855235688 -0.054590905863389085 -2.756558468275941e-06 0.014287625385605138 -0.10206010088995582 0.089574026659087758 -0.046378030282875696 0.076463599471635738 0.088769735219383084 -0.029663948415868675 -0.017902328567800614 0.0021930761031290088 -0.072177998614416775 0.085809604082118407
leaf_weight=0.25187105755321582 0.2352301954815631 0.33727000840008248 1.1983980446821085 0.46755412931088369 0.32063799712341312 0.35872488952008907 0.34889784420374725 0.51386511122109346 0.33710015926044423 0.074520825408399105 0.45200099563226093 0.38386745052412152 0.31427432066993788 0.39580043521709729 0.29991542734205756 0.33226927474606782 0.27595185994869087 0.31000174116343249 0.36615674704080581 0.27630507317371666 0.33381282328628126 0.37104580830782641 0.3290701633086428 0.77214297035243207 0.27769107080530375 0.27412662073038518 0.49186726042535156 0.34880465315655118 0.2882264798390678 0.59754301008069877
leaf_count=23 34 25 123 36 32 32 29 53 22 8 32 24 24 18 24 15 31 19 31 20 32 24 23 37 24 21 32 21 20 111
internal_value=0 -0.0247864 -0.0382745 -0.0502209 0.0055111 0.0171762 0.0217835 0.0269779 0.0168536 0.0239272 0.0119789 0.000829077 0.0144633 -0.00991707 -0.016423 -0.00900304 -0.00185817 0.0129556 0.0230193 -0.0118251 -0.0493695 0.0342371 0.0148986 0.0518667 0.0635097 0.0299354 -0.0479824 0.0424202 -0.0325603 0.0743506
internal_weight=0 2.23845 1.90118 1.66595 9.69649 5.52163 5.26976 4.91103 3.84535 3.50825 2.83619 2.48729 2.03529 4.17486 3.77906 3.45842 3.15851 2.39069 2.08069 0.942387 0.610118 1.43279 1.06174 1.1383 1.06568 0.551818 0.767819 0.732672 0.602501 0.672064
internal_count=1000 218 193 159 782 491 468 436 338 316 197 168 136 291 273 241 217 154 135 67 52 92 68 68 98 45 63 45 44 119
is_linear=0
shrinkage=0.1
Tree=80
num_leaves=31
num_cat=0
split_feature=0 8 7 6 6 25 9 5 22 11 9 4 5 5 7 15 7 2 8 3 7 4 15 6 8 2 4 29 9 4
split_gain=0.146335 0.163441 0.377904 0.251922 0.239951 0.23506 0.218084 0.214001 0.178082 0.244872 0.302114 0.431285 0.278247 0.268535 0.29526 0.22255 0.221625 0.214642 0.270696 0.392448 0.20693 0.124238 0.129407 0.122199 0.0958487 0.0722635 0.0605797 0.0871008 0.040559 0.0303632
threshold=-3.1315989494323726 -0.82658091187477101 -1.6642212867736814 1.7042843699455263 -4.5184216499328604 -0.78593450784683216 -1.9490652680397031 0.56105643510818493 1.4963601827621462 0.10654639080166818 -0.90855789184570301 -2.1543085575103755 0.32781475782394415 -0.67728805541992176 -2.499995112419128 1.1039848923683169 -2.2234120368957515 -1.2476434707641599 1.4360643625259402 0.12777689099311831 -1.4436311721801756 0.69948077201843273 -0.32141780853271479 -0.68222475051879872 0.7093111276626588 3.5917694568634038 0.10320323705673219 -0.13915843516588208 1.3687919378280642 0.4314530491828919
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=29 2 5 8 -4 -2 -5 21 9 10 15 -12 28 14 -11 -3 -15 -16 19 -19 -14 22 -7 -18 -25 -8 -6 -28 -13 -1
right_child=1 3 4 6 26 7 25 -9 -10 13 11 12 20 16 17 -17 23 18 -20 -21 -22 -23 -24 24 -26 -27 27 -29 -30 -31
leaf_value=-0.030771050858475965 0.061339693945417918 -0.090483863717380736 -0.090031106314473752 0.024925642616251267 0.10124653716466597 -0.024702908096035717 -0.089350587580211888 -0.090000733755347603 0.047154585429597884 0.061118354022958656 0.095506891156420401 -0.09558738843403608 -0.043769253765334665 -0.02424261519084402 0.076741615020744836 0.022316875587171658 0.099440648685286667 -0.081697721048163058 -0.090308095121284016 0.048328639632622294 0.061549266698282194 -0.041982266819836703 0.097822908132038658 0.074359414332737281 -0.00076838170520231869 -0.02578210065356102 0.0064038844227428757 0.09647474397271083 -0.040673321086241006 -0.10099940753770038
leaf_weight=0.099568104371429267 0.34934569743927557 0.85660446860129036 0.089295779354870208 0.31161474878899753 0.60396418059826718 0.1556624692166223 0.65842188854003303 0.41014068387448788 0.77825874858535815 0.46949268598109295 0.37148166971746821 0.39634265436325022 0.33834310597740114 0.36387998657301057 0.17578471079468538 0.21978166280314326 0.55372789059765626 0.39893681881949505 0.63489832414779812 0.55512740858830512 0.41585384425707161 0.33710858854465187 0.19316756189800832 0.3339283618843204 0.34554381016641855 0.24550812062807381 0.21152474195696414 0.21802438516169786 0.20358691154979169 0.16128920076880604
leaf_count=17 23 79 4 23 156 13 65 22 31 25 33 36 28 18 9 14 25 25 73 36 31 27 30 24 19 16 23 21 23 31
internal_value=0 0.00157252 0.0237176 -0.00501993 0.0672406 -0.0100911 -0.0472158 -0.0328577 0.00190046 -0.00340902 -0.025876 5.77293e-05 -0.0261271 0.013022 -0.00936135 -0.0674517 0.0443352 -0.0281117 -0.0397115 -0.00604118 0.0143019 0.00130961 0.0431468 0.0645704 0.0361534 -0.0720853 0.0808289 0.0521208 -0.0769522 -0.0741935
internal_weight=0 11.1954 2.56823 8.62712 1.12281 1.44543 1.21554 1.09608 7.41157 6.63331 2.80199 1.72561 1.35413 3.83132 2.23424 1.07639 1.59708 1.76475 1.58896 0.954064 0.754197 0.685939 0.34883 1.2332 0.679472 0.90393 1.03351 0.429549 0.59993 0.260857
internal_count=1000 952 319 633 204 115 104 92 529 498 244 151 118 254 168 93 86 143 134 61 59 70 43 68 43 81 200 44 59 48
is_linear=0
shrinkage=0.1
Tree=81
num_leaves=31
num_cat=0
split_feature=3 2 6 8 2 19 24 27 7 4 23 10 11 2 1 8 20 17 16 5 19 8 8 0 13 2 9 5 12 4
split_gain=0.137145 0.167517 0.181407 0.165687 0.2513 0.191716 0.211864 0.205215 0.175901 0.171814 0.164334 0.162785 0.153655 0.207212 0.178348 0.172446 0.192325 0.224914 0.171204 0.22358 0.152527 0.148125 0.132609 0.119946 0.11357 0.102469 0.0989394 0.0915924 0.0734104 0.0570741
threshold=-1.9173253774642942 4.188784360885621 2.2315721511840825 1.1326235532760622 3.4516642093658452 0.82704880833625805 0.56038278341293346 -1.113314092159271 0.25074130296707159 -1.8600368499755857 0.32948938012123113 -1.5789780020713804 -0.97970148921012867 1.1303530335426333 -1.7068958878517149 -1.7798881530761717 0.44893689453601843 -0.54957136511802662 -0.23371669650077817 -0.57292872667312611 1.228871285915375 -0.82658091187477101 0.72140043973922741 -2.6550524234771724 -0.080186594277620302 0.90473783016204845 -1.9159966707229612 0.20410734415054324 -0.64615306258201588 0.72157731652259838
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=10 2 3 4 8 6 7 -5 12 -9 11 25 13 14 -2 -14 17 -17 -19 -20 -7 -16 23 -10 -4 -1 -26 -23 -8 -18
right_child=1 -3 24 5 -6 20 28 9 22 -11 -12 -13 15 -15 21 16 29 18 19 -21 -22 27 -24 -25 26 -27 -28 -29 -30 -31
leaf_value=-0.020307364599673645 0.07412096325465975 0.067968721724004755 -0.08116956002356418 0.047728978570547656 -0.05596677596818278 0.062123333736842305 0.059944995314150741 -0.0058997164318841432 -0.050989539642246166 -0.082944277954901044 -0.080851255098973895 -0.057819847534476027 0.084659884730194229 -0.08298229839304791 0.037724947086697269 -0.064568254788703511 0.087238888873398659 0.070489359657953929 -0.062034167104577514 0.052534829950702903 -0.027877736163611611 -0.0086248667843465826 -0.023167941182502393 0.090957768085924054 0.065884048707649329 0.06956421068197037 -0.046144619391870728 -0.094249846468226234 -0.011081323067593997 0.029786569211725358
leaf_weight=0.22523011278826743 0.29526987500139501 0.39232821692712683 0.41199983668047935 0.19861195352859828 0.44848675263347093 0.5014276427682488 0.27882429899182171 0.40603000356350127 0.065084534231573343 1.0081198018160646 0.43401546182576567 0.353012160863728 0.44047153776045567 0.32742392568616185 0.30926736589753989 0.43310401367488927 0.36410544108366583 0.42258938378654098 0.37779820605646497 0.31018062448130213 0.30153454758692533 0.25655660429037752 0.15300175012089312 0.69747976117650967 0.12905360746663064 0.29049335769377649 0.20258352108066979 0.24349519156385213 0.30436983122490335 0.32928864425048232
leaf_count=23 24 34 36 14 27 27 14 37 6 105 49 34 47 30 26 27 24 26 25 17 26 20 12 152 15 18 38 21 26 20
internal_value=0 0.00397286 0.00124859 0.00540507 0.0157561 -0.0134876 -0.0287767 -0.047455 0.0221573 -0.0608233 -0.0306027 -0.00549879 0.0133263 -0.0131142 0.00759616 0.0274674 0.0162064 -0.00344492 0.0203922 -0.0103798 0.0283254 -0.0166746 0.0617955 0.0788426 -0.0461077 0.0303149 -0.00254967 -0.0503191 0.0228763 0.0599551
internal_weight=0 9.60849 9.21616 8.47252 5.4736 2.99892 2.19596 1.61276 5.02512 1.41415 1.30275 0.868736 4.10955 1.43201 1.10459 2.67754 2.23707 1.54367 1.11057 0.687979 0.802962 0.809319 0.915566 0.762564 0.743637 0.515723 0.331637 0.500052 0.583194 0.693394
internal_count=1000 876 842 753 504 249 196 156 477 142 124 75 307 121 91 186 139 95 68 42 53 67 170 158 89 41 53 41 40 44
is_linear=0
shrinkage=0.1
Tree=82
num_leaves=31
num_cat=0
split_feature=31 12 8 18 18 31 31 3 29 0 7 1 12 8 21 29 17 27 25 30 9 20 11 8 6 22 4 11 2 1
split_gain=0.134849 0.230469 0.178449 0.173877 0.204064 0.211844 0.251215 0.238853 0.203301 0.166252 0.166027 0.197511 0.162036 0.204449 0.154507 0.154109 0.148104 0.274545 0.140711 0.117191 0.11693 0.111983 0.102942 0.0958586 0.090207 0.0711358 0.0675757 0.0634494 0.0574394 0.100541
threshold=1.2021329402923586 0.18951624631881717 0.84256586432456981 -1.4502596855163572 -0.79747301340103138 -0.1249653398990631 0.18357837945222857 -0.49797442555427546 -0.052107142284512513 0.42796576023101812 -0.27817767858505243 -0.1547611057758331 0.20101968199014666 0.51063084602355968 0.45119327306747442 0.9611392915248872 -0.69839975237846363 -0.59541326761245716 -0.80188676714897145 -0.65799501538276661 -0.39009538292884821 -0.24445675313472745 2.3344395160675053 -0.5930877923965453 0.62837016582489025 -0.40177513659000391 -0.92380249500274647 3.9142535924911503 2.0050599575042729 -0.31010645627975458
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=3 2 14 20 22 7 23 8 24 10 11 -9 28 -14 -2 -10 19 -18 -19 -8 -1 -12 26 -7 -6 -3 -5 -20 29 -11
right_child=1 25 -4 4 5 6 16 9 15 12 21 -13 13 -15 -16 -17 17 18 27 -21 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.018714491216070366 -0.027268698081334961 0.092915726409688579 -0.065723768151836739 -0.027607250699852145 -0.020903525733009739 -0.013902406229972639 -0.026038717753963565 -0.088297860691425992 -0.096337918232188499 0.089953436177080404 -0.024167662126765365 0.020937848917503079 0.069888194665354314 -0.071840868439820169 0.084845868090279455 0.0068077655064752891 0.059133314176868401 0.019013116092576235 -0.077267975874911202 0.061167506985178102 0.068321110687583328 0.075813880267307676 -0.006890940904719773 -0.088270709163505434 0.05331891593797005 0.033621730690196659 -0.092540103315976285 -0.015372176002359697 0.10274008831991788 0.0048042371295830251
leaf_weight=0.22212077095173299 0.24918180634267695 0.41706103755859658 0.34567029151367024 0.23867519636405632 0.32041995890904218 0.23855047317920253 0.22693106345832348 0.31334688636707142 0.43350475956685641 0.33104614238254515 0.16255445388378575 0.35087181435665116 0.20714918681187555 0.200096997898072 0.24258884321898222 0.21754355565644801 0.32629915914731134 0.27132972341496531 0.71150390661205021 0.4801360912970275 0.50598538806661864 0.36038398044183872 0.37843394832452759 0.63387351494748401 0.33488190680509433 0.39298634068109095 0.48791447811526978 0.21586388861760497 0.42880393756786361 0.23862635125988352
leaf_count=17 27 20 35 33 26 28 19 32 37 24 22 29 33 16 24 21 42 27 84 41 44 39 35 60 28 31 52 18 25 31
internal_value=0 0.0261206 -0.0106646 -0.005042 -0.00924533 -0.00294407 -0.0224352 0.0125744 -0.0222944 0.0301421 0.00258785 -0.0305944 0.0534121 0.000250808 0.028037 -0.0618725 -0.00465088 -0.0221907 -0.0443281 0.0331789 0.0417695 0.0447348 -0.0491832 -0.0679359 0.0170267 0.0641498 -0.0712105 -0.0628605 0.0750949 0.0542859
internal_weight=0 1.64749 0.837441 8.83685 8.10874 7.00372 3.10449 3.89923 1.30635 2.59288 1.18716 0.664219 1.40572 0.407246 0.491771 0.651048 2.23206 1.525 1.1987 0.707067 0.728106 0.522938 1.10502 0.872424 0.655302 0.810047 0.72659 0.927368 0.998476 0.569672
internal_count=1000 137 86 863 802 682 319 363 112 251 122 61 129 49 51 58 231 171 129 60 61 61 120 88 54 51 85 102 80 55
is_linear=0
shrinkage=0.1
Tree=83
num_leaves=31
num_cat=0
split_feature=0 1 10 8 7 1 6 11 6 0 28 11 4 1 5 0 19 4 5 4 12 20 21 22 25 8 23 24 8 27
split_gain=0.130332 0.184824 0.236433 0.21844 0.23234 0.204257 0.18262 0.187133 0.23422 0.22976 0.277462 0.176905 0.165965 0.16403 0.155026 0.165973 0.145395 0.139965 0.12545 0.151534 0.122703 0.129994 0.112426 0.101458 0.0876573 0.0668479 0.0423658 0.0525771 0.0344232 0.0323863
threshold=-1.608339190483093 -2.3026347160339351 -1.3095270991325376 -0.82658091187477101 -1.6642212867736814 4.2822875976562509 1.7042843699455263 3.7105647325515752 -2.8922876119613643 0.8787231743335725 -0.87715819478034962 6.7042803764343271 2.6697212457656865 1.6457651853561404 0.049624711275100715 0.58521056175231945 0.66233554482460033 -0.97413113713264454 0.40193679928779608 -0.35809123516082758 -0.55168533325195301 -0.086629480123519884 0.79399952292442333 -0.78451043367385853 -0.80188676714897145 1.7986661791801455 -0.23341625183820722 -0.21857128292322156 0.43258935213088995 -1.712755501270294
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=17 2 25 4 14 29 7 8 -5 10 -10 26 13 -9 15 -3 18 -1 19 -12 21 -11 24 -22 -19 -2 -8 -28 -25 -6
right_child=1 3 -4 6 5 -7 11 12 9 20 16 -13 -14 -15 -16 -17 -18 22 -20 -21 23 -23 -24 28 -26 -27 27 -29 -30 -31
leaf_value=0.03469668972343385 0.099993777282710194 -0.016680936401473542 -0.016818013723444485 0.051564357036643554 -0.010950333798069882 -0.067231919797717338 -0.10205261319054984 0.085582769011539703 -0.076364717265025542 -0.062012974825741009 0.065498214494659507 0.032076070886508962 -0.036755755201412034 -0.00060869424525260079 -0.061605056445179308 0.087470198482576547 0.068434221703735018 -0.005465880414414182 -0.055536301827443325 -0.027960254163903572 -0.013597548822018799 0.034175091713398893 0.008859244702814106 -0.05698881045315838 -0.091504422442449884 0.039331429564853811 -0.089479998079884923 -0.019359890525296287 -0.10165245751586985 0.095293934739245856
leaf_weight=0.24749149824492633 0.43896070460323244 0.38351104082539678 0.45008308457909041 0.4798150565475211 0.030047062435187399 0.093643794534727931 0.3094256492331624 0.56033808540087171 0.36533904599491362 0.27505757834296674 0.3791396731976403 0.20047917822375894 0.275898318272084 0.36437900341115892 0.30706648761406541 0.25457130471477285 0.44879842468071729 0.15450076066190377 0.28963522065896541 0.31984709959943436 0.2390253716148435 0.2872113799676298 0.23702488280832767 0.25028121971990902 0.50696948729455538 0.30990347592160106 0.20830307807773352 0.21973402297589917 0.55569879186805338 0.63586920822853954
leaf_count=22 36 38 38 39 6 4 31 33 32 21 30 16 16 30 20 31 29 19 28 30 18 20 24 23 87 15 18 25 50 171
internal_value=0 0.00393024 0.0404628 -0.00173379 0.0298719 0.0710538 -0.0106712 -0.00320017 -0.0138522 -0.0230568 0.00145269 -0.0512186 0.0313111 0.0516196 -0.00322359 0.0248715 0.021231 -0.0318917 -0.000197551 0.0227329 -0.0505472 -0.0128794 -0.0502335 -0.0708145 -0.0714082 0.0748898 -0.0738623 -0.0534836 -0.087783 0.0905
internal_weight=0 8.93206 1.19895 7.73312 1.70471 0.75956 6.02841 5.09046 3.88985 3.41003 1.80276 0.937942 1.20062 0.924717 0.945149 0.638082 1.43742 1.14599 0.988622 0.698987 1.60727 0.562269 0.898495 1.04501 0.66147 0.748864 0.737463 0.428037 0.80598 0.665916
internal_count=1000 848 89 759 270 181 489 399 320 281 149 90 79 63 89 69 117 152 88 60 132 41 130 91 106 51 74 43 73 177
is_linear=0
shrinkage=0.1
Tree=84
num_leaves=31
num_cat=0
split_feature=8 7 3 4 9 11 16 2 1 27 22 17 24 11 30 8 19 17 9 19 12 30 13 8 2 14 6 19 11 19
split_gain=0.124451 0.214083 0.192318 0.177307 0.239982 0.185382 0.172152 0.202963 0.18798 0.222352 0.172281 0.161349 0.167746 0.134215 0.154662 0.135813 0.120024 0.117936 0.113272 0.0991327 0.116078 0.147053 0.108742 0.101852 0.0982808 0.0906586 0.0883003 0.0836589 0.0835647 0.0710893
threshold=0.93145474791526806 -0.20904585719108579 -2.7369303703308101 -0.16142201423645017 -0.71943753957748402 -0.20719208568334577 0.80493414402008068 3.0034747123718266 0.93677455186843883 -0.88941597938537587 1.001728415489197 0.72848483920097362 -0.2691770046949386 -0.97970148921012867 -1.209358751773834 -1.7129650115966795 0.31949073076248174 0.42055553197860723 -2.0710529088973995 -0.84238478541374195 0.67950376868247997 1.2294318079948428 0.096015840768814101 -0.043257951736450188 0.017128437757492069 0.95362117886543285 -0.60408353805541981 -0.54270070791244496 2.3015670776367192 0.55610954761505138
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 2 -1 4 5 -4 7 8 9 -2 11 12 -11 24 -15 -16 -14 28 -3 -20 21 22 23 -21 -5 -6 -10 -8 -17 -24
right_child=6 18 3 13 25 -7 27 -9 26 10 -12 -13 16 14 15 17 -18 -19 19 20 -22 -23 29 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.088799139060468199 0.062068589582190005 0.10154224440926995 -0.05930471911706852 -0.0074644680878265049 0.086515208126236379 0.033751155518892288 0.00096553621762629895 0.062831081778724873 -0.022694469355447368 -0.071229701536505111 0.039907059395330034 -0.096815876501871193 -0.030682427447083816 -0.075959014011548739 0.055188599538076734 -0.072850704866929711 0.05654942370180438 0.030966167884157011 0.081798204548364672 0.10128762182278402 0.070595261841929241 -0.098596816178072214 0.019240545283511906 -0.019503182421063324 -0.10129574536435114 0.0035188958028706968 -0.10114636583760046 -0.082563895093385839 -0.007215929869934249 -0.074065125686123362
leaf_weight=0.24404723464976985 0.44668429845478508 0.25459178679739114 0.38935925276018679 0.18874511402100314 0.60630828875582599 0.47555744095006958 0.15991624887101352 0.38952695671468962 0.26611297624185681 0.34234927373472701 0.40251009538769711 0.31838701653759927 0.29696106642950248 0.25483163772150863 0.44423162163002405 0.30792578822001926 0.3364243449177593 0.47454919654410332 0.27766721755324386 0.19576344624510944 0.27301007998175919 0.13853301561903208 0.17296488679130562 0.10849639377556741 0.27321172307711095 0.16810015542432666 0.31129462644457806 0.47921611898345862 0.5242023013997823 0.15467966589494608
leaf_count=33 31 47 35 18 72 36 20 20 22 40 25 32 29 13 38 20 25 32 49 54 44 14 34 13 39 25 36 53 35 16
internal_value=0 0.00888395 -0.00255329 0.0025716 0.0280641 -0.00813978 -0.0143949 -0.00468142 -0.0143472 -0.000704133 -0.0172308 -0.0350023 -0.0148322 -0.0143634 -0.00317114 0.00742259 0.015651 -0.00881644 0.0404661 0.0286962 0.0145654 -0.00528915 0.0151668 0.0582147 -0.0629584 0.0684993 -0.0649898 -0.0616641 -0.0315038 -0.0248087
internal_weight=0 5.92678 4.35107 4.10702 1.63933 0.864917 3.74938 3.11025 2.72072 2.14332 1.69663 1.29412 0.975735 2.4677 2.00574 1.75091 0.633385 1.30668 1.57571 1.32111 1.04345 0.770437 0.631904 0.30426 0.461957 0.774408 0.577408 0.639132 0.832128 0.327645
internal_count=1000 667 396 363 168 71 333 260 240 182 151 126 94 195 138 125 54 87 271 224 175 131 117 67 57 97 58 73 55 50
is_linear=0
shrinkage=0.1
Tree=85
num_leaves=31
num_cat=0
split_feature=5 3 5 1 11 20 7 5 12 2 9 4 24 9 6 20 26 18 18 2 12 6 15 3 19 2 11 17 8 9
split_gain=0.119174 0.135615 0.234921 0.191707 0.153494 0.182047 0.162296 0.222621 0.216627 0.153233 0.150708 0.185342 0.163801 0.153278 0.183164 0.135183 0.1321 0.161494 0.140026 0.120821 0.114683 0.177947 0.125991 0.099318 0.0788443 0.0773106 0.0747062 0.0734892 0.0651798 0.0476765
threshold=-4.0636310577392569 2.6622943878173833 -0.67728805541992176 2.8490761518478398 4.6643924713134775 1.9680437445640566 0.72382399439811718 -0.15890216827392575 -0.51321107149124134 2.3993561267852788 -2.8404707908630367 -3.3705158233642574 -0.40476341545581812 1.2301043272018435 -2.4812057018280025 -0.64407771825790394 0.36348965764045721 0.56845283508300792 -0.72996917366981495 0.96589478850364696 -0.15595536679029462 0.51976311206817638 0.34851495921611791 0.36379647254943853 0.8163239359855653 -1.2902153730392454 1.0000000180025095e-35 0.30233138799667364 -0.5797320008277892 -0.92759951949119557
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=-1 3 15 4 5 6 10 8 -8 -6 -2 29 20 16 -15 -3 17 18 -14 -16 -13 22 -22 -5 28 -9 -20 -17 -18 -12
right_child=1 2 -4 23 9 -7 7 25 -10 -11 11 12 13 14 19 27 24 -19 26 -21 21 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.070926621369901161 -0.093245696576062076 0.063324639741856606 0.089549726416568401 -0.099540598094714675 -0.0011893066344759538 0.09140326619659922 0.05318258440117915 -0.006059382834932554 -0.078111954458706695 0.075800684214505976 0.021919929093719599 -0.07556537853935269 -0.040394904810788362 -0.061413305620755081 0.014011146917533697 0.0048006208848277686 -0.015063956949294702 -0.055921848409843447 0.0058754646935271163 0.091183856484452125 -0.036246637168942955 -0.072469467293476517 0.062681979484736666 -0.01941965054793204 0.0066705548938438544 0.08341366882990503 0.071030365458773573 -0.088296341659556354 -0.095433122171704576 0.10121963077968076
leaf_weight=0.23197690473171029 0.23191844031680098 0.17085726582445204 0.45301508266129609 0.28345264797098924 0.51253183931112267 0.19452493020799 0.21985329937888309 0.11877549267956056 0.29333505019894801 0.52160621760413051 0.14049838529899716 0.62313046184135601 0.27150792896281661 0.18037221278063953 0.43565467267762858 0.15654251552768983 0.13843925483524799 0.37820411811117083 0.30215736443643304 0.3796658021165058 0.23425714520271579 0.37192009261343628 0.28578929323703051 0.34065469092456624 0.16068287985399365 0.51662442220549543 0.4214174123480916 0.18499383458402008 0.37223921879194666 0.16468199231894409
leaf_count=36 27 18 65 28 32 12 27 12 37 34 15 68 26 30 35 20 25 39 29 29 36 44 19 38 12 89 27 24 39 28
internal_value=0 0.00166037 0.037087 -0.00256518 0.00188334 -0.00386292 -0.0068322 0.0271231 -0.0218645 0.0376435 -0.0144906 -0.0107329 -0.0157871 -0.00240106 0.0297744 -0.00929618 -0.0180697 -0.00029758 0.0208437 0.0499477 -0.0426488 -0.0196533 0.0181192 -0.0558084 -0.0544231 0.0666884 0.0438224 -0.0456255 -0.0736459 0.0647118
internal_weight=0 9.0593 0.965409 8.0939 7.46979 6.43565 6.24112 1.14859 0.513188 1.03414 5.09254 4.86062 4.55544 3.04034 0.995693 0.512394 2.04465 1.37329 0.995083 0.81532 1.5151 0.891967 0.520046 0.624107 0.671361 0.6354 0.723575 0.341536 0.510678 0.30518
internal_count=1000 964 127 837 771 705 693 165 64 66 528 501 458 291 94 62 197 121 82 64 167 99 55 66 76 101 56 44 64 43
is_linear=0
shrinkage=0.1
Tree=86
num_leaves=31
num_cat=0
split_feature=8 4 9 3 1 0 0 9 26 1 13 19 27 3 9 8 16 2 5 14 15 5 0 10 14 21 6 15 9 27
split_gain=0.112747 0.193647 0.169053 0.158806 0.150909 0.174679 0.172853 0.150649 0.143815 0.154023 0.163788 0.150291 0.141426 0.136889 0.15333 0.20992 0.126 0.148385 0.154734 0.15289 0.145719 0.116059 0.154937 0.108761 0.108706 0.105373 0.0992536 0.0879983 0.0878549 0.0822526
threshold=0.93145474791526806 -0.16142201423645017 -1.4426879882812498 -2.4266439676284786 1.2805632352828982 -0.641687512397766 0.30448183417320257 0.26169002056121832 -0.35152202844619745 0.77400022745132457 -0.79340738058090199 -0.14345091581344602 -1.0820637941360471 2.1290374994277959 -0.2785354852676391 -0.82658091187477101 0.81611305475234996 2.7762584686279301 -0.93297564983367909 0.35118725895881658 0.14212975651025775 -1.2285000085830686 2.1040328741073613 2.4017443656921391 -0.40227904915809626 0.51193365454673778 -3.4843884706497188 0.24113584309816363 1.0000000180025095e-35 -0.41094288229942316
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 13 25 -4 5 12 -6 27 11 10 -10 -9 -2 14 15 23 17 18 20 24 -7 -17 -23 -1 -20 -3 -16 -5 -18 -12
right_child=4 2 3 7 6 16 -8 8 9 -11 29 -13 -14 -15 26 21 28 -19 19 -21 -22 22 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.10151824610292511 0.024461301667596275 0.10149095590350042 -0.1008599259965884 -0.090863353546393175 0.012064260632955025 -0.082043112501756527 -0.10112246357291926 -0.014232651876744665 0.074385692659751063 -0.074288428005532009 0.027070529871922152 0.06919484376588482 -0.10077007693625784 0.090423337338102733 -0.027762681704620529 -0.088862169105311598 -0.095255318164682801 0.083096324664822599 0.041500175064472004 0.079533277352086262 0.010873914872509598 -0.046247311837564364 0.059450192320739663 -0.019926289699939551 -0.037741927872576131 0.0018164829035358843 0.067156289459481003 -0.0077037391417275024 0.00011118143082661881 -0.058714656499896295
leaf_weight=0.23349438476725526 0.12843770300969481 0.2098189789685424 0.20745210675522763 0.31434673583134998 0.21036278409883369 0.31737104454077902 0.37623215978965163 0.38129649619804695 0.23295601073186911 0.36527567682787776 0.17516436055302687 0.49789103906368914 0.30273028576630245 0.32729417242808256 0.12977504852460697 0.23870020359754607 0.18973403243580822 0.28062037983909238 0.33466189308092098 0.37848097551614046 0.36050080065615475 0.3661220357753332 0.22324725170619786 0.10778186804964207 0.35863495711237026 0.21448219491867349 0.72901240266219258 0.21378942253068089 0.19679213454946876 0.30883015535073111
leaf_count=47 16 29 36 24 20 35 46 34 24 30 15 37 68 89 15 18 27 15 20 20 25 41 17 17 26 17 118 26 15 33
internal_value=0 0.00876658 -0.00756813 -0.0167989 -0.0143448 -0.00483168 -0.0605318 -0.00979421 0.00297051 -0.0214357 0.00549205 0.033013 -0.0634658 0.0304126 0.0207282 -0.00283512 0.00562892 0.0155916 0.0047648 0.0284152 -0.0326287 -0.0300355 -0.00621014 0.0631636 0.000509093 0.051106 0.0528127 -0.0572004 -0.0467014 -0.0276678
internal_weight=0 5.47673 3.1213 2.697 3.43456 2.84796 0.586595 2.48955 1.96141 1.08223 0.716951 0.879188 0.431168 2.35543 2.02813 1.16935 2.4168 2.03027 1.74965 1.07178 0.677872 0.828069 0.589369 0.341276 0.693297 0.424301 0.858787 0.528136 0.386526 0.483995
internal_count=1000 667 305 259 333 267 66 223 173 102 72 71 84 362 273 140 183 141 126 66 60 76 58 64 46 46 133 50 42 48
is_linear=0
shrinkage=0.1
Tree=87
num_leaves=31
num_cat=0
split_feature=6 0 15 8 4 9 26 21 0 1 10 20 8 30 18 6 28 27 30 21 3 7 22 1 25 23 10 2 12 8
split_gain=0.111012 0.143898 0.171782 0.146942 0.122074 0.294466 0.153286 0.150565 0.147177 0.166059 0.181309 0.266148 0.177688 0.169973 0.166036 0.200477 0.149366 0.137582 0.120902 0.113369 0.105124 0.0965233 0.0928983 0.0800679 0.0684102 0.0651229 0.0632622 0.0585266 0.0885774 0.0479995
threshold=2.2315721511840825 0.46007883548736578 -0.8999110460281371 0.11012277007102968 -1.0175390839576719 -2.2834039926528926 -0.66544929146766651 0.60416230559349071 -1.5799034237861631 0.3707870244979859 -0.40247164666652674 0.060031592845916755 -1.6528002023696897 0.42361395061016088 0.31057116389274603 -0.56465166807174672 -0.68067798018455494 -0.27425691485404963 -1.2146136760711668 0.79399952292442333 1.823546886444092 -2.577299833297729 -0.33938346803188318 1.6292831897735598 -0.80188676714897145 -0.85004675388336171 -3.0313315391540523 2.0050599575042729 -0.14266175031661985 0.69922864437103283
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=4 -2 -3 -4 5 7 22 29 19 12 14 21 -10 16 -11 -16 -14 -15 -8 24 27 -12 -7 -24 -6 -20 -27 28 -18 -1
right_child=1 2 3 -5 8 6 18 -9 9 10 11 -13 13 17 15 -17 20 -19 25 -21 -22 -23 23 -25 -26 26 -28 -29 -30 -31
leaf_value=-0.030805456364942453 -0.08776672142392955 -0.091159691961066783 0.08859943278833457 -0.045133999256959621 -0.0092958255545747704 0.04590367773955964 -0.021341018468976625 0.037081629462908994 0.10133264191487693 -0.082877384688447173 -0.087455461161163969 0.086100551280600957 -0.0246777013757211 0.0053604041859369725 0.070265649463287852 -0.066730648068955692 -0.0012386037305286842 -0.083344886384781131 0.0083901577103449743 0.018291523472593038 0.0022684478889105964 0.010218823519943182 0.0053786058211993081 -0.097864184508235391 -0.089742877526446593 0.016880670316873743 0.083154059168347727 0.10193013427658099 0.079637721317690549 -0.10104747969550065
leaf_weight=0.1575240510865118 0.31811740022385482 0.14965710858814418 0.21305448855855502 0.13373378512915224 0.141745274013374 0.18635586387244985 0.18359030404826659 0.17256291233934462 0.23124768357956771 0.6025883024849461 0.19557084876578301 0.31067128898575902 0.41511006688233454 0.427577495604055 0.20409433369059116 0.2241164474398829 0.2320214946521445 0.29581801290623833 0.17577311757486347 0.20112119481200352 0.39704626752063632 0.20961460485705163 0.15727206721203479 0.14379931535222568 0.41575298173120234 0.16812293877592344 1.0052623582596405 0.28112293686717749 0.32525254576467055 0.25439373636618257
leaf_count=20 55 11 47 13 19 25 24 19 31 61 15 32 35 37 18 23 22 28 26 20 33 33 32 13 83 25 121 19 25 35
internal_value=0 -0.0352609 -0.00161575 0.0370271 0.00355574 0.0211753 0.0392609 -0.041335 -0.00542443 0.00166097 -0.0221957 0.0164659 0.0176557 0.0095047 -0.049045 -0.00143531 0.0272191 -0.0309138 0.0547947 -0.0460701 0.0446565 -0.0369256 -0.009586 -0.0439328 -0.069289 0.065155 0.0736584 0.0647305 0.0459648 -0.0741858
internal_weight=0 0.814563 0.496445 0.346788 7.71513 2.60466 2.02018 0.584481 5.11047 4.35185 1.74666 0.715857 2.6052 2.37395 1.0308 0.428211 1.65055 0.723396 1.53275 0.758619 1.23544 0.405185 0.487427 0.301071 0.557498 1.34916 1.17339 0.838397 0.557274 0.411918
internal_count=1000 126 71 60 874 340 266 74 534 412 182 80 230 199 102 41 134 65 196 122 99 48 70 45 102 172 146 66 47 55
is_linear=0
shrinkage=0.1
Tree=88
num_leaves=31
num_cat=0
split_feature=18 3 8 4 11 8 6 30 31 12 8 24 4 31 17 22 7 4 3 4 8 7 29 8 22 9 7 27 29 9
split_gain=0.10513 0.121639 0.153991 0.139635 0.217271 0.151825 0.221888 0.14932 0.233786 0.221911 0.223926 0.164257 0.153091 0.132608 0.104891 0.097846 0.0944811 0.0934168 0.137315 0.0930666 0.0929096 0.0885395 0.0840689 0.0821581 0.0797625 0.0651277 0.0934698 0.0746272 0.0639357 0.0176874
threshold=-1.5012420415878294 -1.9173253774642942 2.9477462768554692 3.1555353403091435 4.5497055053710946 -2.4568742513656612 2.2315721511840825 0.55339282751083385 1.3691098093986513 0.30352285504341131 -1.0830669999122617 -0.59483751654624928 -3.1882129907608028 -0.70689350366592396 0.23587644100189212 -0.38123951852321619 -1.1933641433715818 -1.1025146245956419 0.46754512190818792 0.61790910363197338 -0.92787450551986683 0.25074130296707159 0.54209741950035106 1.0356987714767458 0.26117700338363653 1.2301043272018435 0.84199178218841564 -0.52868872880935658 -0.73016235232353199 -1.4426879882812498
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=17 20 3 4 5 28 7 8 9 11 -11 14 -9 -14 -7 16 -13 -1 -19 -4 -2 -12 -15 29 -22 26 27 -17 -3 -6
right_child=1 2 19 -5 23 6 -8 12 -10 10 21 15 13 22 -16 25 -18 18 -20 -21 24 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.079399514209179275 0.019874338729586216 0.0038783011114899321 -0.082271276460531048 -0.048554465429207237 0.10151532895130488 0.011109610429591299 -0.090905679090791433 0.072528913594149313 0.085732571906235233 0.062865830216948068 -0.088595258079126257 -0.065702416779634107 0.019047874950758409 -0.085572016762863273 -0.080502520911519521 0.061102274515735466 0.037631646327678868 -0.053657815273244897 0.062537978650335488 0.0075547382593429109 -0.088641328732591451 -0.006673536408106283 -0.0052059789481493169 0.0098148980753929443 -0.019922000699902132 0.092624904418135368 0.095063523251671916 -0.016019286636761162 0.096141711294585513 0.064829307291514066
leaf_weight=0.29728620836976916 0.20918758411426097 0.1112932562828064 0.33741959300823499 0.44277332338970143 0.23264800710603595 0.30661333963507731 0.2695914166397414 0.12957592913880944 0.39373273131786835 0.16908483018050902 0.47997496195603151 0.15052076894789934 0.25078302569454536 0.52661032957257681 0.21096991939703003 0.21356148098129701 0.21468198765069191 0.19922591472277407 0.20776750613003969 0.17524914914974943 0.35955229878891248 0.18193660024553537 0.17290011013392359 0.23466015700250864 0.31854552146978676 0.32147610827814788 0.20869877230143175 0.30418832344003344 0.23100333561887865 0.30204004887491454
leaf_count=28 32 29 44 38 26 32 42 22 36 27 57 16 32 62 22 16 32 18 22 16 52 29 33 20 32 30 27 31 69 28
internal_value=0 -0.00363248 0.00105967 0.0055122 0.00977447 0.00193869 -0.00293979 0.00265953 0.0136438 0.00336623 -0.0398413 0.0219632 -0.0294372 -0.0433407 -0.026232 0.0396155 -0.00495818 0.0367861 0.00565939 -0.0515654 -0.0383865 -0.0660778 -0.0657077 0.059143 -0.0563595 0.0551495 0.0385655 0.0157918 0.0661435 0.0807918
internal_weight=0 7.45927 6.57199 6.05932 5.61655 4.8472 4.5049 4.23531 3.15544 2.76171 0.830996 1.93071 1.07987 0.950293 0.517583 1.41313 0.365203 0.70428 0.406993 0.512669 0.887285 0.661912 0.69951 0.769348 0.678098 1.04792 0.726449 0.51775 0.342297 0.534688
internal_count=1000 932 816 756 718 644 546 504 355 319 113 206 149 127 54 152 48 68 40 60 116 86 95 74 84 104 74 47 98 54
is_linear=0
shrinkage=0.1
Tree=89
num_leaves=31
num_cat=0
split_feature=1 6 8 0 15 11 8 3 22 27 11 11 1 18 19 13 16 4 9 21 23 26 4 2 4 10 26 1 9 15
split_gain=0.105125 0.111709 0.128119 0.193164 0.214958 0.123153 0.11008 0.151462 0.157997 0.168354 0.138119 0.12088 0.106671 0.155919 0.153204 0.182173 0.149422 0.142716 0.103816 0.103058 0.0958936 0.109146 0.109675 0.0868862 0.0868698 0.0792194 0.0748299 0.0610281 0.0554128 0.0442572
threshold=-3.7169742584228511 1.7042843699455263 -0.82658091187477101 -0.016390442848205563 -1.2552280426025388 6.7042803764343271 0.93145474791526806 2.0606999397277836 0.26117700338363653 0.88477838039398204 -4.6548492908477774 0.95515820384025585 0.30255699157714849 -0.064929060637950883 0.82704880833625805 0.52293393015861522 -0.1172221079468727 -1.2004670500755308 0.84960302710533153 0.59227067232131969 -1.12335729598999 -0.76153719425201405 -0.29410147666931147 1.4999915361404421 -0.25337916612625117 -2.6730151176452632 -0.32592111825942988 0.46325501799583441 0.16526672244071963 0.30211952328681951
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=25 6 3 -3 -5 -4 7 8 9 11 -8 24 14 -14 15 17 -15 -12 -9 20 -10 -22 -23 -20 27 -1 -13 -2 -11 -19
right_child=1 2 5 4 -6 -7 10 18 19 28 12 26 13 16 -16 -17 -18 29 23 -21 21 22 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.10205921392660239 0.022754723285362061 -0.06661458507214367 -0.075688611184085999 -0.090777897350730288 0.086310189161653256 0.02502114767880112 -0.10110511380384953 0.10124349794759269 0.10201989347846874 0.00081105579096517962 0.031710711605421854 0.04572283847069937 -0.085334656262979647 0.045280147987903807 0.066264081564474872 0.04266913385779679 -0.058177342199236259 -0.10131087763191926 0.075816590722757649 0.071341999639912543 -0.07225866274432112 0.070903789884343607 -0.02800285380900391 -0.020982777734194713 -0.086630066084890225 0.0051171024506322556 -0.031952119252851546 -0.055545048345785546 0.077294900631891345 -0.04036611286433394
leaf_weight=0.17173177079530433 0.17457201145589341 0.23631653864867985 0.61172463177354053 0.087834425561595708 0.31211462318606209 0.15149360103532672 0.16490880266064811 0.2991653379285707 0.097533594846026972 0.1693370412394869 0.19248937023803625 0.25182684604078531 0.35293464665301144 0.3145807912806049 0.3665939439088105 0.42637013248167932 0.25097884814022098 0.23422654447494995 0.16761503909947351 0.43893551762448624 0.17242079309653491 0.22380738644278608 0.22464781993767224 0.20754192202002741 0.42329207819420922 0.16556466254405666 0.24438894793274812 0.23160707383067347 0.21499148168368265 0.24253698455868289
leaf_count=13 35 32 86 6 74 16 24 63 11 22 16 23 51 24 26 35 28 33 31 36 21 39 27 24 60 27 28 35 26 28
internal_value=0 -0.00259678 -0.0280724 0.00506568 0.0474192 -0.0556984 0.00326013 0.0146609 0.00463231 -0.0146791 -0.0126016 -0.0315735 -0.00647106 -0.0331788 0.0103055 -0.00841818 -0.000631264 -0.0409651 0.0573046 0.0331656 0.00984046 -0.00464003 0.0213578 0.0222659 -0.0549291 0.0544744 0.0074675 -0.0218925 0.0435957 -0.0703073
internal_weight=0 7.48679 1.39948 0.636266 0.399949 0.763218 6.0873 3.54168 2.86736 1.71002 2.54562 1.32569 2.38071 0.918494 1.46222 1.09562 0.56556 0.669253 0.674322 1.15735 0.71841 0.620876 0.448455 0.375157 0.829471 0.337296 0.496216 0.406179 0.384329 0.476764
internal_count=1000 960 214 112 80 102 746 481 363 229 265 181 241 103 138 112 52 77 118 134 98 87 66 55 130 40 51 70 48 61
is_linear=0
shrinkage=0.1
Tree=90
num_leaves=31
num_cat=0
split_feature=5 1 3 7 20 11 5 25 13 13 18 4 7 8 0 27 27 14 8 7 0 29 25 24 19 0 6 24 24 7
split_gain=0.100151 0.155668 0.16336 0.141796 0.161648 0.171832 0.167429 0.136999 0.131378 0.138171 0.151117 0.120575 0.123075 0.113598 0.166159 0.110067 0.0985827 0.104839 0.0895956 0.114719 0.115807 0.0845026 0.0801803 0.0748621 0.0713789 0.0693701 0.0692627 0.0683819 0.0647415 0.0561436
threshold=-2.8203281164169307 0.43324142694473272 2.6009738445281987 0.72382399439811718 0.47881735861301428 0.95515820384025585 -0.77069473266601551 -0.63787114620208729 1.0069346427917483 -1.2598152160644529 0.23372912406921389 1.4291652441024782 -1.5343946218490598 -0.11924156546592711 0.088062286376953139 0.50145037472248088 -0.080120045691728578 -0.5278467833995818 0.81758242845535289 -2.0654069185256954 1.3665547370910647 0.74896043539047252 0.11230562999844552 -0.20663078874349591 0.16267675906419757 -1.0424582958221433 -0.13115194439887998 1.0000000180025095e-35 -0.52980807423591603 -1.1751348376274107
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=11 7 3 4 18 6 -6 21 9 -9 13 12 -1 22 -15 -16 17 -5 19 -3 27 25 -11 -12 -25 -2 -10 -21 -4 -27
right_child=1 2 28 16 5 -7 -8 8 26 10 23 -13 -14 14 15 -17 -18 -19 -20 20 -22 -23 -24 24 -26 29 -28 -29 -30 -31
leaf_value=0.041111073769295237 -0.000547518084568799 -0.078600098744865954 -0.0052425966506058283 -0.090180623945832983 0.056189300677484434 0.038816292022438602 -0.098246269774960399 0.05733520749420136 0.0081220535451929608 -0.011709854342954788 -0.082950074963699855 -0.10113293434127785 -0.054603761480779736 -0.087100957450151964 -0.023448918742703398 0.06319912857854118 0.090675384021430896 0.035725110120207779 -0.1009259296271463 0.078986353386883479 -0.064192468826363924 -0.010124514360081896 0.07412040268166506 -0.052673982045708612 0.039656117513152851 0.10181244260170075 0.086090663840383622 -0.019173704134750221 0.097715470442017976 0.046896119237580275
leaf_weight=0.2494982642820105 0.15951474476605654 0.41799305361928496 0.095750188920646906 0.099096316611394286 0.098735918756574392 0.48909160954644904 0.24289489738293912 0.31038843342685218 0.19713100127410144 0.16607872811437208 0.36441962566459551 0.20105543307727203 0.29106576746562485 0.23814157384913348 0.36451186635531485 0.24523041141219437 0.16249293870350801 0.19883080580621026 0.32900654373224825 0.15452148951590094 0.20790159108582884 0.21692879352485761 0.31579630250053015 0.22719195560784999 0.13259927317267284 0.3437866362510249 0.26997008093167085 0.13125134393339977 0.16864738886943076 0.4060400717717132
leaf_count=30 23 52 18 9 13 41 38 35 22 25 42 38 52 39 35 15 39 38 50 16 23 24 49 34 29 26 33 25 37 50
internal_value=0 0.00369607 -0.0143656 -0.0221765 -0.0333339 0.000805452 -0.0536123 0.016457 0.00472287 -0.00485138 -0.0142488 -0.0350172 -0.0104264 0.00576854 -0.0162657 0.0113998 0.0280196 -0.00615357 -0.0561927 -0.0400491 -0.00740813 0.0459568 0.0445389 -0.0510036 -0.0186462 0.0593354 0.0531855 0.0339029 0.0604297 0.0720746
internal_weight=0 6.75394 2.79621 2.53182 2.0714 0.830722 0.341631 3.95773 2.83146 2.36436 2.05397 0.741619 0.540564 1.32976 0.847884 0.609742 0.46042 0.297927 1.24067 0.911667 0.493674 1.12627 0.481875 0.724211 0.359791 0.909341 0.467101 0.285773 0.264398 0.749827
internal_count=1000 880 399 344 258 92 51 481 358 303 268 120 82 163 89 50 86 47 166 116 64 123 74 105 63 99 55 41 55 76
is_linear=0
shrinkage=0.1
Tree=91
num_leaves=31
num_cat=0
split_feature=0 1 10 8 7 1 3 29 6 10 6 8 4 11 15 9 5 0 2 22 25 4 24 30 9 3 7 8 31 14
split_gain=0.0961295 0.126917 0.16274 0.136605 0.177375 0.124962 0.12446 0.183264 0.216606 0.12709 0.115383 0.150683 0.128279 0.142333 0.166335 0.10014 0.0940983 0.112449 0.0881521 0.119566 0.0943221 0.0744638 0.0701012 0.0596519 0.0517228 0.0440158 0.028542 0.0275024 0.0211504 0.0209947
threshold=-1.608339190483093 -2.3026347160339351 -1.3095270991325376 -0.82658091187477101 -1.6642212867736814 4.2822875976562509 -0.84586244821548451 -0.62766262888908375 -1.008602976799011 -4.4574854373931876 0.77008655667305004 1.0356987714767458 -0.67423412203788746 3.7105647325515752 0.56999582052230846 -2.0369137525558467 0.049624711275100715 0.58521056175231945 1.4031839370727541 0.88484010100364696 -0.80188676714897145 -0.72163537144660939 -0.19467142969369886 -0.043422264978289597 0.59585255384445202 1.399665415287018 -0.50514426827430714 1.4360643625259402 -1.8477005362510679 -0.67656695842742909
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=18 2 24 4 16 28 7 23 -9 -8 12 22 15 14 25 -11 17 -3 19 20 -1 -22 -12 -5 -2 -14 -13 -17 -6 -10
right_child=1 3 -4 6 5 -7 9 8 29 10 11 26 13 -15 -16 27 -18 -19 -20 -21 21 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.026091882139394978 0.038399937178539741 -0.018131121977673647 -0.014254640750721173 -0.019637720191090214 0.015488009421118887 -0.064843654160944952 -0.06060259966875188 0.018866735443598591 -0.059596977032825731 -0.0034154872008087902 -0.032896518692602024 -0.034491638904455765 -0.094076806224653794 0.048851653160184583 0.040862406505502324 0.095156200440221772 -0.059570944004979524 0.08403069676245184 -0.084616692825179335 0.066503419344784542 0.0089123007490042785 -0.10085304428288003 0.043248159215367246 0.057313260042593905 0.10137466625459209 -0.033204773764043342 -0.10089134922009742 0.049084485362377868 0.095957719359407084 -0.10098915296729101
leaf_weight=0.15598221053369343 0.23003073642030358 0.28701086912769824 0.34305111803405441 0.1936104791238904 0.035381011519348249 0.058962106239050627 0.29768136679194857 0.24867069907486439 0.16298613173421472 0.21963448880705982 0.21344599733129144 0.094877082854509354 0.21268571773544065 0.3920421241200529 0.24691847991198301 0.32314311672234897 0.21788923279382288 0.17249148662085645 0.23066306285909377 0.14702919917181134 0.08641374867875129 0.21701334996032517 0.2788705895654856 0.21001010597683489 0.30118675797712058 0.26906460663303733 0.20378381147747859 0.21629695990122855 0.42514230425877009 0.49377962309517831
leaf_count=17 22 38 38 23 4 4 30 22 24 28 20 14 25 28 24 46 20 31 56 16 13 50 30 19 29 27 41 17 173 71
internal_value=0 0.0040534 0.0394339 -0.00159693 0.0282663 0.072226 -0.00995287 -0.0356394 -0.0606208 0.00137465 0.00828259 -0.0237598 0.0217654 0.000266408 -0.0258737 0.0535069 -0.00544606 0.0202192 -0.031999 -0.0119855 -0.0371051 -0.0695927 0.0102353 0.0204011 0.074105 -0.0600789 -0.0797978 0.076683 0.0897754 -0.0907171
internal_weight=0 6.34865 0.874269 5.47438 1.19688 0.519485 4.2775 1.30906 0.905436 2.96844 2.67076 0.790977 1.87979 1.12071 0.728669 0.759075 0.677392 0.459502 0.837102 0.606439 0.459409 0.303427 0.492317 0.403621 0.531217 0.48175 0.298661 0.53944 0.460523 0.656766
internal_count=1000 848 89 759 270 181 489 159 117 330 300 105 195 104 76 91 89 69 152 96 80 63 50 42 51 52 55 63 177 95
is_linear=0
shrinkage=0.1
Tree=92
num_leaves=31
num_cat=0
split_feature=8 4 26 11 9 7 8 10 1 5 0 1 6 7 13 22 5 14 27 24 4 14 27 2 9 19 8 29 8 23
split_gain=0.0890226 0.167154 0.1603 0.122743 0.211769 0.171165 0.158197 0.179912 0.142662 0.136722 0.172928 0.133195 0.120242 0.120304 0.113672 0.102791 0.0813425 0.0875199 0.0766828 0.0761736 0.074664 0.0737882 0.0722392 0.0687178 0.0662934 0.0504373 0.0418922 0.0514689 0.0272715 0.0237561
threshold=2.4709751605987553 -3.4333997964859004 -0.19718474894762036 -0.97970148921012867 -1.9159966707229612 2.0273846387863164 -1.7129650115966795 4.5763163566589364 -2.1899367570877071 -1.9025785923004148 2.4160362482070927 -0.80352920293807972 0.22959581017494204 -1.8709572553634641 -1.0066322088241575 0.26117700338363653 -0.63844007253646839 0.75826820731163036 0.52713641524314891 0.083792209625244155 1.5504462718963625 0.25446638464927679 1.310772240161896 2.1345607042312627 -2.6924784183502193 -0.34891957044601435 -0.94087243080139149 -0.39794003963470453 -0.2367703318595886 -0.062162378802895539
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 24 -2 5 18 11 20 8 -8 10 23 15 14 -14 -11 19 17 -4 -5 28 -6 25 26 -10 -1 -16 27 -13 -3 -15
right_child=2 3 16 4 6 -7 7 -9 9 12 -12 22 13 29 21 -17 -18 -19 -20 -21 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.0050649422373392839 -0.10078859659860523 -0.032031394639245335 -0.10085610641221071 0.085346537417147608 0.10075377942579818 0.091532579038749298 0.059688097816935265 -0.090837661644462803 0.0084581989220458408 -0.064518777644870559 -0.056816056739217316 -0.094334671766402312 -0.0012432610557940177 -0.099615645525664959 -0.049280260306459288 0.045366023899635878 0.031561448244974151 0.013310389413902612 0.012807012017939309 0.015866922256463821 0.016661217483331114 0.056996262703547532 0.012767863298272918 0.083232618254617002 0.088042189167686691 0.027283913586509913 -0.091841358246813609 -0.0015281610532221043 -0.099744833075133363 -0.056779428922637924
leaf_weight=0.10116131289396435 0.23527133886818774 0.10757484484929591 0.13340770278591663 0.43402810549014192 0.1980347130447625 0.11946092855941959 0.33013883669627864 0.24837826058501367 0.19545447808923277 0.2057385506923316 0.1854333789087832 0.10247837411588989 0.28644132940098643 0.30846686687436886 0.17113362869713511 0.2833083551377058 0.31408417341299344 0.13519405038096011 0.21939390804618597 0.18504282995127141 0.22616442246362567 0.3011079098796472 0.11401466641109437 0.33110388077329844 0.31334072156460013 0.17303717532195106 0.43599680450279266 0.14334225776838128 0.13303187378915016 0.22310217632912099
leaf_count=11 43 19 27 63 62 40 34 39 34 17 23 14 29 29 21 35 25 14 23 24 26 29 19 32 84 20 81 33 24 26
internal_value=0 0.00402025 -0.0311205 -0.000467756 0.00887178 -0.0236817 -0.0011927 -0.00937874 -0.00191601 -0.0104577 0.026231 -0.0328282 -0.0261087 -0.0534857 0.000206397 -0.00130668 -0.00299054 -0.0433932 0.0609906 -0.0323716 0.0559193 0.0208432 -0.0609087 0.0554769 0.0653189 -0.0107864 -0.0732291 -0.0402176 -0.0694703 -0.0816371
internal_weight=0 6.07591 0.817957 5.66141 4.03716 1.62425 3.38374 2.95954 2.71116 2.38102 0.711992 1.50479 1.66903 0.81801 0.851017 0.708958 0.582686 0.268602 0.653422 0.42565 0.424199 0.645279 0.795832 0.526558 0.414502 0.344171 0.681817 0.245821 0.240607 0.531569
internal_count=1000 891 109 796 507 289 421 333 294 260 89 249 171 84 87 102 66 41 86 67 88 70 147 66 95 41 128 47 43 55
is_linear=0
shrinkage=0.1
Tree=93
num_leaves=31
num_cat=0
split_feature=6 0 2 30 8 3 27 3 1 26 30 11 5 19 0 30 17 25 7 1 8 21 4 10 21 19 31 14 0 11
split_gain=0.0818133 0.111276 0.129509 0.0831503 0.0901749 0.166078 0.127873 0.1105 0.125006 0.100176 0.101794 0.0959882 0.124587 0.0915558 0.114538 0.114009 0.0901055 0.0874214 0.0818159 0.144213 0.0801585 0.0731135 0.116914 0.0684596 0.0597772 0.055835 0.0542331 0.0715123 0.0848986 0.0416111
threshold=1.7042843699455263 0.06608057022094728 1.4812332987785342 -1.8685542941093443 0.93145474791526806 2.0606999397277836 -1.4307758212089536 -2.7369303703308101 2.5908429622650151 -1.0634013414382932 1.1657342910766604 -3.1654969453811641 -2.1957887411117549 0.60753285884857189 -0.70722478628158558 0.99595102667808544 -0.34909805655479426 -0.032054329290986054 0.88489025831222545 0.71218520402908336 -0.75641953945159901 -0.4913369864225387 -0.44345408678054804 4.310413360595704 0.59807986021041881 -0.40020120143890375 0.45617610216140753 -0.44417506456375117 0.77826052904129039 -0.38942228257656092
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=3 -2 24 -1 5 6 -5 -8 13 -6 11 25 16 14 17 26 -13 -9 19 21 -4 -14 -23 -7 -3 -11 27 -16 -29 -15
right_child=1 2 20 4 9 23 7 8 -10 10 -12 12 18 29 15 -17 -18 -19 -20 -21 -22 22 -24 -25 -26 -27 -28 28 -30 -31
leaf_value=-0.068793773250343235 -0.072021279459926965 -0.012213162988722716 0.074297636991585786 -0.072355101008265824 -0.076442260448627022 0.091047079722581359 -0.07538868788505157 -0.061657025072584332 -0.044564775789149268 0.0010770809580244504 -0.068498338934068786 0.01665655405807856 0.080684588724576567 0.026696470183374267 0.094028031642763882 -0.043536509143871908 -0.10071378701520446 0.022795714553190474 0.07808307481594437 -0.054370838545324908 -0.005549748270547418 0.070087179897329793 -0.022457071843885405 0.011920312295065849 -0.095553389168806888 -0.083908808636930413 0.086169381808570769 -0.046347415591464873 0.05299690000755735 0.092080466159984553
leaf_weight=0.14811857946915541 0.34169116456177995 0.17942982277600095 0.2113288372347597 0.18520420376444269 0.1943390738742895 0.38864557891793072 0.12997342424932945 0.26340627850731835 0.26536276793922287 0.14308972470462322 0.20384903188096348 0.12098928389605135 0.249747655238025 0.14931620369316079 0.13347618861007571 0.14664232631912455 0.14238280578865703 0.22924887234694391 0.26606016675941657 0.22783247998449951 0.31038675468880683 0.23221530334558349 0.3312287047738196 0.15214755424676696 0.1654005419841269 0.16815203492296915 0.29719476812169887 0.15754083381034323 0.18949302355758846 0.27958925534039725
leaf_count=18 72 39 49 28 33 83 20 38 32 15 32 19 14 42 19 26 35 34 20 31 37 22 34 35 21 30 43 25 29 25
internal_value=0 -0.0236927 -0.00463602 0.00509568 0.00718147 0.0186726 0.00750371 0.0141028 0.0196121 -0.00777406 -0.0013753 0.0058963 0.015951 0.028838 0.0165852 0.0373412 -0.0467956 -0.0223583 0.0285942 0.0159461 0.0267936 0.0356468 0.0156837 0.0687854 -0.0521879 -0.0448375 0.0525913 0.0318233 0.00789817 0.0693181
internal_weight=0 1.20824 0.866546 5.39525 5.24713 2.96724 2.42645 2.24124 2.11127 2.27989 2.08555 1.8817 1.57046 1.84591 1.417 0.924347 0.263372 0.492655 1.30708 1.04102 0.521716 0.813192 0.563444 0.540793 0.34483 0.311242 0.777705 0.48051 0.347034 0.428905
internal_count=1000 218 146 782 764 479 361 333 313 285 252 220 175 281 214 142 54 72 121 101 86 70 56 118 60 45 116 73 54 67
is_linear=0
shrinkage=0.1
Tree=94
num_leaves=31
num_cat=0
split_feature=1 4 8 0 3 11 6 31 12 24 7 0 29 7 10 5 7 4 21 17 25 15 9 10 9 25 25 31 4 26
split_gain=0.080568 0.0833191 0.157415 0.104131 0.103009 0.175385 0.125071 0.105625 0.171439 0.172447 0.148218 0.156665 0.102527 0.11845 0.0930351 0.0907174 0.10755 0.0978194 0.089605 0.0862314 0.079545 0.0779754 0.0665605 0.0611396 0.0545293 0.0531501 0.0500422 0.0455708 0.0373315 0.0364464
threshold=-3.7169742584228511 3.2311422824859624 -0.95728960633277882 0.18048211932182315 -1.9173253774642942 4.8875966072082528 2.2129242420196538 1.3691098093986513 0.079420387744903578 1.1405742764472964 0.25074130296707159 0.77826052904129039 -0.5291345715522765 -2.1114803552627559 -4.0208497047424308 -0.95051786303520192 -1.2958345413208006 -0.61726528406143177 0.18265433609485629 -0.36764360964298243 -0.67077535390853871 0.076843108981847777 0.71947535872459423 -2.6730151176452632 1.3687919378280642 -1.0000000180025095e-35 -0.72430333495140065 0.71584683656692516 -0.16142201423645017 -0.25410135090351099
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=23 2 3 12 14 6 7 8 9 15 24 -12 28 -14 -4 17 22 -6 21 -9 -3 -18 -17 -1 -10 -7 -16 -5 -2 -19
right_child=1 20 4 27 5 25 -8 19 10 -11 11 -13 13 -15 26 16 18 29 -20 -21 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.10158457540118355 -0.0048847917654098912 0.0055849378512151382 0.015038344580540257 0.09371178085755473 0.010523006861176229 0.018884405590873635 -0.1007425084646329 -0.014223153155154093 -0.089188305414537733 0.082636732817972833 0.055747389317953727 -0.060557321649691402 -0.01790051435450677 0.084112575152337693 -0.021968267102219594 -0.06999914339102252 -0.055744520758087036 -0.030588740130604364 0.078188838093781035 0.074641232831470269 -0.087030883274620052 0.048335006376369796 0.01093914394048124 0.006967818742194952 -0.022044431147331744 0.084018470823482239 -0.10076720823528272 0.023741501324776242 -0.084399486471108745 -0.10152321413191095
leaf_weight=0.13289538555545732 0.12103002492222004 0.16067048395052552 0.15916641161311418 0.30205578241293551 0.22672454372514039 0.2394146332226228 0.14623508960357856 0.18355196033371612 0.51678078589611665 0.25278637508745294 0.2414122762857005 0.22262119315564632 0.20994072797475383 0.24860444942896698 0.11712274985620752 0.1955012164544313 0.13526566495420633 0.11382269518799148 0.29183240362908691 0.26956135570071627 0.21932076534721989 0.15386009000940248 0.21154528320766985 0.14049448369769368 0.15791291056666523 0.26280088291969139 0.25839718588394944 0.13453997533360962 0.11528798379004002 0.19919671516981896
leaf_count=13 25 14 22 124 33 27 34 23 79 24 36 35 33 50 15 25 19 14 32 24 30 26 19 27 37 26 50 38 17 29
internal_value=0 -0.00253472 0.000494519 0.0338781 -0.007797 -0.00231345 -0.0102038 -0.00627783 -0.0132502 0.00612975 -0.0435531 -5.00805e-05 0.00983109 0.0374068 -0.0490331 -0.00652935 0.0114815 -0.0394983 0.0390984 0.0386431 -0.0478704 -0.000357945 -0.0279349 0.0529612 -0.0734732 0.052968 -0.0761902 0.07215 -0.0436761 -0.0757294
internal_weight=0 6.06696 5.68697 1.13146 4.55551 4.02083 3.51861 3.37238 2.91926 1.78053 1.13873 0.464033 0.694863 0.458545 0.534686 1.52775 0.988005 0.539744 0.580958 0.453113 0.379991 0.289126 0.407046 0.27339 0.674694 0.502216 0.37552 0.436596 0.236318 0.313019
internal_count=1000 960 916 287 629 542 489 455 408 221 187 71 125 83 87 197 121 76 77 47 44 45 44 40 116 53 65 162 42 43
is_linear=0
shrinkage=0.1
Tree=95
num_leaves=31
num_cat=0
split_feature=8 4 26 11 9 6 7 2 9 4 8 30 6 1 5 3 0 30 31 2 18 27 9 6 19 3 8 7 8 10
split_gain=0.0811211 0.146089 0.13694 0.104373 0.176565 0.131123 0.126119 0.135566 0.157355 0.122411 0.160028 0.135192 0.106839 0.0956808 0.180015 0.194079 0.157618 0.08791 0.0925558 0.0846953 0.0683752 0.0634313 0.059585 0.0516727 0.0430808 0.0414132 0.0373788 0.0309322 0.0240788 0.00510644
threshold=2.4709751605987553 -3.4333997964859004 -0.19718474894762036 -0.97970148921012867 -1.9159966707229612 2.2315721511840825 -0.88453030586242665 0.29434275627136236 1.4186322093009951 0.50270438194274913 1.5303894281387331 -0.54829272627830494 -1.5804541110992429 0.86920762062072765 -0.42181968688964838 1.3817856311798098 -0.78594365715980519 0.40916936099529272 0.14827194809913638 -1.6019626855850218 -0.2674479186534881 0.52713641524314891 -1.1903105974197385 -1.2840110063552854 -0.25734533369541163 -1.3282699584960935 -1.2448169589042661 -1.8011893033981321 1.3702826499938967 -1.5154966711997984
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 22 -2 6 21 9 7 8 -3 10 11 -6 -11 14 15 -14 -15 18 19 -8 -4 28 -1 -19 -21 -13 -7 -18 -5 -9
right_child=2 3 20 4 5 26 17 29 -10 12 -12 25 13 16 -16 -17 27 23 -20 24 -22 -23 -24 -25 -26 -27 -28 -29 -30 -31
leaf_value=0.020049150125201993 -0.10070539775101561 -0.060718523849459909 -0.050982144580117933 0.1002407634342225 -0.0086738200047520794 -0.027487692113458945 -0.098503616382257256 -0.073704641643919391 0.075423751335871536 -0.078838330642892598 -0.045421072655663905 0.02838146415518792 0.036822008945711783 0.049716555435968805 0.077524627253259135 -0.070851035122604794 -0.10134254133441062 0.010646088047800088 0.087032401081566083 0.090382958978887373 0.024471989702253127 0.012779391103262198 0.10079990487154399 -0.081079189691536491 -0.0067668116324164978 0.094656734864074998 -0.10079488671084134 -0.040972997805440781 0.045755138100881297 -0.10088607123966697
leaf_weight=0.16515052353497595 0.20543254664517008 0.22015723545337096 0.19300174695672467 0.26219649575068604 0.24424173528677784 0.13161016869707964 0.070632342831231654 0.09142761689145118 0.13818453764542937 0.20255071786232326 0.2056828904314898 0.11995595484040678 0.36067523551173519 0.14517160202376544 0.32447019557002854 0.31240060209529474 0.19122112856712192 0.10983158109593205 0.18781470944668399 0.072530739504145458 0.317934986000182 0.1913792816339992 0.20456528235808935 0.13932410176494159 0.12314385354693491 0.44054660836991499 0.1475186084571759 0.15261079045012593 0.11743817548267543 0.28320804610848427
leaf_count=30 43 30 27 51 39 26 9 19 22 35 28 12 32 10 36 38 22 22 42 11 39 23 65 39 38 84 34 25 12 57
internal_value=0 0.00407739 -0.0317536 -0.000420028 0.00878327 -0.000981708 -0.0231658 -0.0521921 -0.00821904 0.00576494 0.0332972 0.0534166 -0.010705 -0.00142142 0.0163414 -0.0131532 -0.0376568 0.00708635 0.0332745 -0.00463859 -0.00403013 0.0597216 0.0647289 -0.0406453 0.0292437 0.0804728 -0.0662303 -0.0745473 0.0833859 -0.0942526
internal_weight=0 5.35564 0.716369 4.98592 3.54967 2.97866 1.43625 0.732977 0.358342 2.69953 1.01043 0.804744 1.6891 1.48655 0.997546 0.673076 0.489004 0.703277 0.454122 0.266307 0.510937 0.571014 0.369716 0.249156 0.195675 0.560503 0.279129 0.343832 0.379635 0.374636
internal_count=1000 891 109 796 507 421 289 128 52 361 163 135 198 163 106 70 57 161 100 58 66 86 95 61 49 96 60 47 63 76
is_linear=0
shrinkage=0.1
Tree=96
num_leaves=31
num_cat=0
split_feature=0 1 8 25 1 20 11 6 29 3 24 23 6 28 21 30 1 5 5 9 29 25 10 8 14 29 13 18 4 4
split_gain=0.0790786 0.102749 0.0990681 0.111805 0.0862341 0.103463 0.0952063 0.0975559 0.125842 0.0930136 0.106778 0.0784148 0.0900889 0.106463 0.0783054 0.0749067 0.0737857 0.0830662 0.0694386 0.0688711 0.0680174 0.0679091 0.061077 0.0613623 0.0490641 0.0439719 0.036456 0.0344568 0.0679806 0.0343606
threshold=-1.608339190483093 3.992965936660767 -1.6528002023696897 1.2622607946395876 -3.7169742584228511 0.59053277969360363 4.8875966072082528 -0.28123188018798823 -0.047192480415105813 -0.84586244821548451 -0.64478895068168629 0.72400054335594188 -1.6811078190803526 -0.70314431190490712 0.099660724401474013 -0.66628974676132191 1.3164123892784121 0.27409940958023077 -0.85830512642860401 -0.072343081235885606 -0.068975701928138719 -0.64878144860267628 -2.8759547472000118 -0.37335616350173945 0.049458399415016181 -0.34217648208141321 -0.80384975671768177 0.05846847593784333 1.2886417508125307 -0.6461447775363921
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=16 2 3 27 -4 9 7 20 -9 15 21 12 -12 -14 24 -6 17 26 -13 -16 -7 -11 -17 -24 -15 -23 -1 -2 -29 -10
right_child=1 -3 4 -5 5 6 -8 8 29 10 11 18 13 14 19 22 -18 -19 -20 -21 -22 25 23 -25 -26 -27 -28 28 -30 -31
leaf_value=-0.015607704537829174 0.10079352560017721 -0.070384644678079686 0.062021935406102985 -0.041396044375179727 -0.1008510623238015 -0.0033634840787542288 0.080243484434377385 0.03275390031550323 -0.030247495668142877 0.025088676606379985 -0.076020309790370014 0.00074565841142650296 -0.045075212752370808 0.01738715382691338 -0.065154528312023494 0.012370611943426101 0.033707375771193392 -0.0035943272464337567 0.074627465093638157 0.039318946722564072 0.079175907168842027 -0.01168340380438338 0.0015319288532450875 -0.077771955899146983 0.089805425160577046 -0.078603403469830732 -0.1006359749801758 0.079279736623608515 -0.048433012222147603 -0.098870744600996791
leaf_weight=0.063590506790205836 0.20659719448303793 0.17878889688290645 0.22648571367608283 0.11029849317856133 0.16040940704988316 0.16666754760080948 0.23690071236342192 0.25841964795836248 0.15278265468077734 0.13458352023735642 0.14401870343135637 0.21375762589741498 0.25739527607220192 0.16073651635088293 0.12435172568075359 0.26767680095508695 0.13500520482193679 0.23037078278139234 0.31419472728157405 0.12810139858629555 0.24899005342740566 0.14797164424089715 0.15391657550935633 0.26651717093773131 0.22383751068264246 0.29185435370891344 0.24354810506338243 0.14630228183523286 0.058282648795284331 0.13966774992877595
leaf_count=17 70 24 28 21 28 28 24 35 22 18 26 27 35 30 27 38 22 35 32 17 37 21 29 48 27 41 78 83 10 22
internal_value=0 0.00408843 0.00678337 0.0480051 0.0019191 -0.00132756 0.0234306 0.00950551 -0.0180916 -0.0112946 -0.00018931 0.0133349 -0.00261853 0.00920053 0.031131 -0.039313 -0.0323853 -0.0489856 0.0447142 -0.0121419 0.0460798 -0.0370694 -0.0249676 -0.0487396 0.0595375 -0.0560894 -0.0830316 0.0719867 0.0428966 -0.0630204
internal_weight=0 5.11951 4.94072 0.521481 4.41924 4.19275 1.20343 0.966528 0.55087 2.98932 2.1408 1.56639 1.03844 0.894422 0.637027 0.84852 0.672515 0.537509 0.527952 0.252453 0.415658 0.57441 0.688111 0.420434 0.384574 0.439826 0.307139 0.411182 0.204585 0.29245
internal_count=1000 848 824 184 640 612 168 144 79 444 301 221 162 136 101 143 152 130 59 44 65 80 115 77 57 62 95 163 93 44
is_linear=0
shrinkage=0.1
Tree=97
num_leaves=31
num_cat=0
split_feature=8 4 26 3 1 11 26 13 10 2 30 25 4 0 22 25 18 5 15 14 10 9 3 7 9 9 8 9 9 5
split_gain=0.0753376 0.122731 0.117605 0.0901252 0.12829 0.180662 0.117945 0.105491 0.102165 0.101251 0.0855332 0.0994078 0.123474 0.0974637 0.120255 0.0734482 0.0731966 0.0631729 0.0631057 0.0614174 0.0573239 0.106933 0.0555377 0.0512878 0.0504395 0.043981 0.0361472 0.0316434 0.0311959 0.0297543
threshold=2.4709751605987553 -3.4333997964859004 -0.19718474894762036 -0.22925552725791928 0.46325501799583441 0.40085124969482427 -0.051145128905773156 -0.3785655796527862 -2.8759547472000118 4.188784360885621 0.50428083539009105 0.10375598073005678 1.6072534918785097 0.37160789966583258 0.77968162298202526 -0.62472391128540028 -0.064929060637950883 -0.63844007253646839 0.14212975651025775 -0.048332408070564263 0.69830441474914562 -1.1705294847488401 -1.1233239173889158 -2.3758826255798335 -1.1705294847488401 0.2187360227108002 0.14835035800933841 2.3834861516952519 0.68402928113937389 -0.7574045956134795
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=1 20 -2 4 5 8 7 -6 22 10 11 12 15 27 19 25 -17 18 -4 -15 -1 -22 -3 -13 -9 -5 -25 -12 -10 -18
right_child=2 3 17 9 6 -7 -8 24 28 -11 13 23 -14 14 -16 16 29 -19 -20 -21 21 -23 -24 26 -26 -27 -28 -29 -30 -31
leaf_value=0.096885956927485611 -0.10061971248526236 -0.026741087718923019 -0.10083074312920406 0.0095262659771922064 0.046483341567355903 0.065214203836779824 -0.084807338925080772 0.0051496798452265171 -0.099734873505858287 0.094665407173429417 -0.08156647179399136 0.00085378766539251035 -0.050020826280321784 -0.07190818513755097 0.091850605869257654 -0.043386817610343693 0.07542681494774317 0.027281883141839887 0.0048319603729122468 0.024354490558636505 -0.072628662190455232 0.10070487946449934 0.060645946153748014 0.086226329337818042 -0.077915006239692983 0.085560353826817703 0.035538048332741864 -0.014620676296499498 -0.032180039867039455 0.0008622085684549945
leaf_weight=0.19794520361756429 0.18042282856185921 0.15098514303099364 0.10656376043334603 0.12027430681337126 0.19132976210676134 0.35191556619247422 0.39859999797772616 0.1150567305739969 0.15280530847667229 0.1383107102010398 0.23670792498160143 0.185845852131024 0.34895981679437682 0.14053756176144816 0.12674567352223676 0.23379963115439739 0.10532558645354573 0.26213402696885157 0.12036784720839933 0.12543551228009164 0.071474173688329756 0.070894134667469189 0.14031264814548194 0.28770421684748715 0.20048580481670797 0.20702284388244152 0.2753239298472181 0.10061742446850985 0.12368944066110998 0.1087949695502185
leaf_count=47 43 27 24 29 26 35 79 13 36 23 51 32 39 30 19 39 19 25 17 23 19 29 18 74 42 28 44 23 26 21
internal_value=0 0.00414581 -0.031615 -0.000171537 -0.0173887 0.00891743 -0.0441084 -0.012103 -0.0259751 0.0112913 0.00686141 0.0181782 -0.000624626 -0.0221736 0.0116897 0.0216108 -0.00470094 -0.00615823 -0.0447857 -0.0265098 0.0620793 0.013685 0.0153516 0.0464041 -0.047627 0.0576195 0.0614395 -0.0615979 -0.0695143 0.0375404
internal_weight=0 4.9069 0.669488 4.56659 1.82518 0.919708 0.905472 0.506872 0.567793 2.74141 2.6031 1.87305 1.12418 0.730044 0.392719 0.775217 0.44792 0.489066 0.226932 0.265973 0.340314 0.142368 0.291298 0.748874 0.315543 0.327297 0.563028 0.337325 0.276495 0.214121
internal_count=1000 891 109 796 302 142 160 81 107 494 471 325 175 146 72 136 79 66 41 53 95 48 45 150 55 57 118 74 62 40
is_linear=0
shrinkage=0.1
Tree=98
num_leaves=31
num_cat=0
split_feature=0 1 5 6 9 7 4 8 11 0 5 22 4 6 28 9 13 4 28 17 8 9 10 6 9 22 22 23 24 3
split_gain=0.0708472 0.0867613 0.108285 0.123066 0.115347 0.106937 0.105617 0.125748 0.126472 0.134057 0.131332 0.139509 0.0973698 0.0824481 0.0808079 0.0752993 0.0681708 0.0630496 0.0631122 0.0634381 0.0601695 0.0586943 0.0551927 0.0517991 0.0606151 0.050617 0.0466239 0.0442533 0.0352543 0.0233148
threshold=-1.608339190483093 -0.33602377772331232 -0.95051786303520192 -0.59109562635421742 0.35674107074737554 -1.0387114286422727 1.5018882751464846 -2.4568742513656612 0.55282053351402294 1.1855481863021853 -0.63424587249755848 -0.59525883197784413 -3.3526372909545894 -1.5970541238784788 -0.60353773832321156 -2.096564412117004 0.37701444327831274 -0.97413113713264454 -0.014420712832361458 0.42055553197860723 1.2666346430778506 1.3436141610145571 -2.1437025070190425 0.82417592406272899 -0.141118049621582 0.62442603707313549 0.48932586610317236 -0.28794039785861963 -0.39458091557025904 -0.29190585017204279
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=17 2 3 -2 14 16 7 -3 12 23 11 -11 15 -6 -4 -9 -5 -1 -19 -20 25 22 -8 -10 -25 28 27 -16 -14 -13
right_child=1 6 4 5 13 -7 21 8 9 10 -12 29 20 -15 26 -17 -18 18 19 -21 -22 -23 -24 24 -26 -27 -28 -29 -30 -31
leaf_value=0.030454226544273894 -0.082015664846297354 0.095493600577092067 -0.046977464552757899 0.0077627663933966806 -0.0092747788152871772 -0.041818308933520516 -0.072127016833293423 -0.036283930886380898 0.079831534449184255 0.044641132264002369 0.041285773465924734 -0.10028094843244356 0.0020608536304149217 0.083834880453037747 0.04707436725245237 0.10057128180066098 0.097732238540230498 -0.088604216625886204 -0.059585013730046869 0.036985025474055411 -0.10063254383761976 -0.089222525346306661 0.017324431530033765 -0.040625210811454211 0.070278449928394007 0.028058244759545676 0.068961150630794193 -0.036779796006906146 -0.067680109770170777 -0.044094179198602312
leaf_weight=0.12848347797989845 0.18223306385334581 0.13421178496355501 0.19277606735704467 0.190187941363547 0.11592398956418049 0.21988692713785019 0.11766046588309109 0.076952330186031759 0.31786449949140638 0.12368551934196215 0.28482158701808591 0.20760813006199963 0.097024855378549546 0.52947880707506545 0.11927157136960886 0.084188180917408317 0.15115077246446162 0.21655140571238007 0.15225884690880764 0.12295850776717998 0.18092888212413527 0.21180237139924429 0.16670902085024852 0.09899708823650144 0.09813485958147794 0.10590031647006981 0.1844811164191924 0.13324384979205184 0.28655837621772695 0.11462937395845008
leaf_count=22 31 77 29 29 17 36 15 13 41 18 35 38 25 93 18 33 18 72 37 21 37 28 29 26 14 26 21 22 54 25
internal_value=0 0.00402499 0.0197167 -0.010616 0.0374015 0.0125678 -0.00767357 0.00168137 -0.00437974 0.0157798 -0.0117594 -0.0456405 -0.0345805 0.067111 0.00695466 0.0352163 0.0476028 -0.0319212 -0.048218 -0.0164405 -0.0513569 -0.0493698 -0.0196869 0.0548559 0.0145841 -0.033143 0.0307462 0.00282736 -0.0500396 -0.0802937
internal_weight=0 4.72631 2.01863 0.743459 1.27518 0.561226 2.70768 2.21151 2.07729 1.24574 0.730745 0.445923 0.831553 0.645403 0.629773 0.161141 0.341339 0.620252 0.491769 0.275217 0.670412 0.496172 0.284369 0.514996 0.197132 0.489484 0.436997 0.252515 0.383583 0.322238
internal_count=1000 848 314 114 200 83 534 462 385 197 116 81 188 110 90 46 47 152 130 58 142 72 44 81 40 105 61 40 79 63
is_linear=0
shrinkage=0.1
Tree=99
num_leaves=31
num_cat=0
split_feature=0 1 8 7 6 6 20 8 3 5 1 0 0 22 11 14 27 27 4 2 4 11 16 5 13 6 31 1 1 4
split_gain=0.0674541 0.0696553 0.0741955 0.166677 0.0899929 0.0941181 0.0965746 0.0829 0.0862338 0.0955048 0.0789118 0.0871315 0.0765944 0.0818907 0.0972902 0.0959447 0.0777583 0.0752579 0.0668864 0.0605962 0.0603125 0.0682629 0.0589837 0.0663064 0.0561723 0.0467039 0.0392907 0.0307815 0.0299616 0.0150091
threshold=-3.1315989494323726 4.1692261695861825 0.93145474791526806 -0.86519092321395863 1.9243513345718386 1.1906984448432925 0.94570809602737438 -1.0830669999122617 1.5328662991523745 0.97322815656661998 1.2805632352828982 0.30448183417320257 -0.641687512397766 0.81031483411788952 3.7105647325515752 1.0592966675758364 -1.0820637941360471 0.27165855467319494 -1.5903732776641843 2.3993561267852788 0.61790910363197338 -0.67806494235992421 -0.039748009294271462 -0.048709154129028313 -0.49590902030467982 0.059982538223266609 0.38249349594116216 -0.93864721059799183 -0.4237337708473205 -0.86077696084976185
decision_type=2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
left_child=-1 2 3 4 5 6 17 -5 9 18 12 -12 16 14 15 22 -4 20 -9 29 21 -2 23 -14 -20 -19 -7 -24 -22 -6
right_child=1 -3 10 7 19 26 -8 8 -10 -11 11 -13 13 -15 -16 -17 -18 25 24 -21 28 -23 27 -25 -26 -27 -28 -29 -30 -31
leaf_value=-0.075071944951751762 -0.046999072568029904 -0.071918235304448228 0.027641404290291466 0.096883408387115108 -0.049069729509923658 0.095611743769455068 0.065541161312771809 0.045198124799364636 0.090939198951104347 0.069682127965000187 0.011014701594283235 -0.1006276503138893 -0.052627535009820953 0.042960234379214049 0.048489002811411136 0.051486404644959953 -0.10045359624143338 -0.0077087143166597264 0.014580719948671895 0.0079997309260280754 -0.030761626656559096 0.043797026079306106 -0.033322164847490364 0.032712756722109153 -0.075794358131254014 0.0637990789497795 0.019136308309067349 -0.10065789099852285 -0.089449497628592389 -0.098396611902252809
leaf_weight=0.11738886896637257 0.17963373329257515 0.12558365869335819 0.070364486775361001 0.24251844967511715 0.11107260789140128 0.14532132641761553 0.16856457435642358 0.12955580640118569 0.18051554734120145 0.19781767061795108 0.11157737800385792 0.18718020082451403 0.20873764104908332 0.41156410262919951 0.2226585391908884 0.13207910989876837 0.14513787950272661 0.24144141713622957 0.11146845109760761 0.12869615360978059 0.15193568938411772 0.153613223941648 0.10812804859597203 0.16147008491679993 0.17955871479352958 0.14691452111583203 0.12493959249695763 0.18242636477225505 0.20350550662260503 0.13873577141202975
leaf_count=48 28 20 15 148 21 27 33 22 49 37 19 40 33 39 22 19 52 34 22 30 21 27 15 17 32 23 22 38 25 22
internal_value=0 0.00161605 0.00350952 0.0135393 -0.00412745 0.00676375 -0.00484289 0.0456752 0.0301305 0.01238 -0.0116582 -0.0589324 -0.0030598 0.00533174 -0.00991843 -0.0263213 -0.0586288 -0.0158585 -0.0145716 -0.0477455 -0.0357086 -0.00514578 -0.0418743 -0.0154054 -0.0411791 0.0193426 0.0602577 -0.0755993 -0.064363 -0.0764643
internal_weight=0 5.00272 4.87713 2.93581 1.89437 1.51587 1.24561 1.04143 0.798916 0.618401 1.94132 0.298758 1.64257 1.42706 1.0155 0.792841 0.215502 1.07704 0.420583 0.378505 0.688688 0.333247 0.660762 0.370208 0.291027 0.388356 0.270261 0.290554 0.355441 0.249808
internal_count=1000 952 932 623 313 240 191 310 162 113 309 59 250 183 144 122 67 158 76 73 101 55 103 50 54 57 49 53 46 43
is_linear=0
shrinkage=0.1
end of trees
feature_importances:
Column_8=256
Column_4=219
Column_6=205
Column_9=202
Column_0=187
Column_7=179
Column_1=172
Column_3=168
Column_5=146
Column_11=138
Column_2=94
Column_10=86
Column_30=67
Column_13=63
Column_20=60
Column_19=56
Column_25=55
Column_22=53
Column_26=53
Column_15=52
Column_27=52
Column_18=51
Column_29=49
Column_24=45
Column_31=42
Column_12=41
Column_14=38
Column_17=38
Column_21=37
Column_16=36
Column_23=30
Column_28=30
parameters:
[boosting: gbdt]
[objective: binary]
[metric: binary_logloss]
[tree_learner: serial]
[device_type: cpu]
[linear_tree: 0]
[data: ]
[valid: ]
[num_iterations: 100]
[learning_rate: 0.1]
[num_leaves: 31]
[num_threads: 0]
[deterministic: 0]
[force_col_wise: 0]
[force_row_wise: 0]
[histogram_pool_size: -1]
[max_depth: 25]
[min_data_in_leaf: 20]
[min_sum_hessian_in_leaf: 0.001]
[bagging_fraction: 1]
[pos_bagging_fraction: 1]
[neg_bagging_fraction: 1]
[bagging_freq: 0]
[bagging_seed: 3]
[feature_fraction: 1]
[feature_fraction_bynode: 1]
[feature_fraction_seed: 2]
[extra_trees: 0]
[extra_seed: 6]
[early_stopping_round: 0]
[first_metric_only: 0]
[max_delta_step: 0]
[lambda_l1: 0]
[lambda_l2: 0]
[linear_lambda: 0]
[min_gain_to_split: 0]
[drop_rate: 0.1]
[max_drop: 50]
[skip_drop: 0.5]
[xgboost_dart_mode: 0]
[uniform_drop: 0]
[drop_seed: 4]
[top_rate: 0.2]
[other_rate: 0.1]
[min_data_per_group: 100]
[max_cat_threshold: 32]
[cat_l2: 10]
[cat_smooth: 10]
[max_cat_to_onehot: 4]
[top_k: 20]
[monotone_constraints: ]
[monotone_constraints_method: basic]
[monotone_penalty: 0]
[feature_contri: ]
[forcedsplits_filename: ]
[refit_decay_rate: 0.9]
[cegb_tradeoff: 1]
[cegb_penalty_split: 0]
[cegb_penalty_feature_lazy: ]
[cegb_penalty_feature_coupled: ]
[path_smooth: 0]
[interaction_constraints: ]
[verbosity: -1]
[saved_feature_importance_type: 0]
[max_bin: 255]
[max_bin_by_feature: ]
[min_data_in_bin: 3]
[bin_construct_sample_cnt: 200000]
[data_random_seed: 1]
[is_enable_sparse: 1]
[enable_bundle: 1]
[use_missing: 1]
[zero_as_missing: 0]
[feature_pre_filter: 1]
[pre_partition: 0]
[two_round: 0]
[header: 0]
[label_column: ]
[weight_column: ]
[group_column: ]
[ignore_column: ]
[categorical_feature: ]
[forcedbins_filename: ]
[objective_seed: 5]
[num_class: 1]
[is_unbalance: 0]
[scale_pos_weight: 1]
[sigmoid: 1]
[boost_from_average: 1]
[reg_sqrt: 0]
[alpha: 0.9]
[fair_c: 1]
[poisson_max_delta_step: 0.7]
[tweedie_variance_power: 1.5]
[lambdarank_truncation_level: 30]
[lambdarank_norm: 1]
[label_gain: ]
[eval_at: ]
[multi_error_top_k: 1]
[auc_mu_weights: ]
[num_machines: 1]
[local_listen_port: 12400]
[time_out: 120]
[machine_list_filename: ]
[machines: ]
[gpu_platform_id: -1]
[gpu_device_id: -1]
[gpu_use_dp: 0]
[num_gpu: 1]
end of parameters
pandas_categorical:null
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/model_repository | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/model_repository/sk_model/config.pbtxt | name: "sk_model"
backend: "fil"
max_batch_size: 8192
input [
{
name: "input__0"
data_type: TYPE_FP32
dims: [ 32 ]
}
]
output [
{
name: "output__0"
data_type: TYPE_FP32
dims: [ 1 ]
}
]
instance_group [{ kind: KIND_GPU }]
parameters [
{
key: "model_type"
value: { string_value: "treelite_checkpoint" }
},
{
key: "predict_proba"
value: { string_value: "false" }
},
{
key: "output_class"
value: { string_value: "true" }
},
{
key: "threshold"
value: { string_value: "0.5" }
},
{
key: "algo"
value: { string_value: "ALGO_AUTO" }
},
{
key: "storage_type"
value: { string_value: "AUTO" }
},
{
key: "blocks_per_sm"
value: { string_value: "0" }
}
]
dynamic_batching {
preferred_batch_size: [1, 2, 4, 8, 16, 32, 64, 128, 1024, 2048, 4096, 8192]
max_queue_delay_microseconds: 30000
} | 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/.helmignore | # Patterns to ignore when building packages.
# This supports shell glob matching, relative path matching, and
# negation (prefixed with !). Only one pattern per line.
.DS_Store
# Common VCS dirs
.git/
.gitignore
.bzr/
.bzrignore
.hg/
.hgignore
.svn/
# Common backup files
*.swp
*.bak
*.tmp
*~
# Various IDEs
.project
.idea/
*.tmproj
.vscode/
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton/Chart.yaml | # Copyright (c) 2019-2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
apiVersion: v1
appVersion: "2.8"
description: Triton Inference Server
name: triton-inference-server
version: 2.8.7 | 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton/values.yaml | # Copyright (c) 2019-2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
initReplicaCount: 1
minReplicaCount: 2
maxReplicaCount: 6
HPATargetAverageValue: 80
DCGMExporterService: dcgm-exporter
image:
imageName: < update this with the image name you just pushed >
pullPolicy: IfNotPresent
modelRepositoryPath: s3://S3_BUCKET_PATH/model_repository
numGpus: 1
logVerboseLevel: 0
allowGPUMetrics: True
secret:
- region: < replace with base64 AWS_REGION >
- id: < replace with base64 AWS_SECRET_KEY_ID >
- key: < replace with base64 AWS_SECRET_ACCESS_KEY >
- session_token: < replace with base64 AWS_SESSION_TOKEN >
service:
type: NodePort
deployment:
livenessProbe:
failureThreshold: 60
initialDelaySeconds: 10
periodSeconds: 5
successThreshold: 1
timeoutSeconds: 1
readinessProbe:
failureThreshold: 60
initialDelaySeconds: 5
periodSeconds: 5
successThreshold: 1
timeoutSeconds: 1
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton/templates/_helpers.tpl | {{/*
# Copyright (c) 2019-2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/}}
{{/* vim: set filetype=mustache: */}}
{{/*
Create inference server name.
*/}}
{{- define "triton-inference-server.name" -}}
{{- default .Chart.Name .Values.nameOverride | trunc 63 | trimSuffix "-" -}}
{{- end -}}
{{/*
Create a default fully qualified app name.
We truncate at 63 chars because some Kubernetes name fields are limited to this (by the DNS naming spec).
If release name contains chart name it will be used as a full name.
*/}}
{{- define "triton-inference-server.fullname" -}}
{{- if .Values.fullnameOverride -}}
{{- .Values.fullnameOverride | trunc 63 | trimSuffix "-" -}}
{{- else -}}
{{- $name := default .Chart.Name .Values.nameOverride -}}
{{- if contains $name .Release.Name -}}
{{- .Release.Name | trunc 63 | trimSuffix "-" -}}
{{- else -}}
{{- printf "%s-%s" .Release.Name $name | trunc 63 | trimSuffix "-" -}}
{{- end -}}
{{- end -}}
{{- end -}}
{{/*
Create chart name and version as used by the chart label.
*/}}
{{- define "triton-inference-server.chart" -}}
{{- printf "%s-%s" .Chart.Name .Chart.Version | replace "+" "_" | trunc 63 | trimSuffix "-" -}}
{{- end -}}
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton/templates/service.yaml | # Copyright (c) 2019-2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
apiVersion: v1
kind: Service
metadata:
name: {{ template "triton-inference-server.name" . }}
namespace: {{ .Release.Namespace }}
labels:
app: {{ template "triton-inference-server.name" . }}
chart: {{ template "triton-inference-server.chart" . }}
release: {{ .Release.Name }}
heritage: {{ .Release.Service }}
spec:
type: {{ .Values.service.type }}
ports:
- port: 8000
targetPort: http
name: http-inference-server
- port: 8001
targetPort: grpc
name: grpc-inference-server
- port: 8002
targetPort: metrics
name: metrics-inference-server
selector:
app: {{ template "triton-inference-server.name" . }}
release: {{ .Release.Name }} | 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton/templates/hpa.yaml | # Copyright (c) 2019-2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
apiVersion: autoscaling/v2beta1
kind: HorizontalPodAutoscaler
metadata:
name: triton-hpa
namespace: {{ .Release.Namespace }}
labels:
app: triton-hpa
spec:
minReplicas: {{ .Values.minReplicaCount }}
maxReplicas: {{ .Values.maxReplicaCount }}
metrics:
- type: Object
object:
metricName: DCGM_FI_DEV_GPU_UTIL
targetValue: {{ .Values.HPATargetAverageValue }}
target:
kind: Service
name: {{ .Values.DCGMExporterService }}
scaleTargetRef:
apiVersion: apps/v1
kind: Deployment
name: {{ template "triton-inference-server.name" . }} | 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton/templates/secrets.yaml | # Copyright (c) 2019-2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
apiVersion: v1
kind: Secret
metadata:
name: aws-credentials
type: Opaque
data:
AWS_DEFAULT_REGION: {{ .Values.secret.region }}
AWS_ACCESS_KEY_ID: {{ .Values.secret.id }}
AWS_SECRET_ACCESS_KEY: {{ .Values.secret.key }}
AWS_SESSION_TOKEN: {{.Values.secret.session_token }} | 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton/templates/istio-gateway.yaml | # Copyright (c) 2019-2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
apiVersion: networking.istio.io/v1alpha3
kind: Gateway
metadata:
name: triton-gateway
spec:
selector:
istio: ingressgateway # use istio default controller
servers:
- port:
number: 80
name: http
protocol: HTTP
hosts:
- "*"
- port:
number: 31400
name: grpc
protocol: TCP
hosts:
- "*" | 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton/templates/deployment.yaml | # Copyright (c) 2019-2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
apiVersion: apps/v1
kind: Deployment
metadata:
name: {{ template "triton-inference-server.name" . }}
namespace: {{ .Release.Namespace }}
labels:
app: {{ template "triton-inference-server.name" . }}
chart: {{ template "triton-inference-server.chart" . }}
release: {{ .Release.Name }}
heritage: {{ .Release.Service }}
spec:
replicas: {{ .Values.initReplicaCount }}
selector:
matchLabels:
app: {{ template "triton-inference-server.name" . }}
release: {{ .Release.Name }}
template:
metadata:
labels:
app: {{ template "triton-inference-server.name" . }}
release: {{ .Release.Name }}
spec:
containers:
- name: {{ .Chart.Name }}
image: "{{ .Values.image.imageName }}"
imagePullPolicy: {{ .Values.image.pullPolicy }}
resources:
limits:
nvidia.com/gpu: {{ .Values.image.numGpus }}
args: ["tritonserver",
"--model-store={{ .Values.image.modelRepositoryPath }}",
"--model-control-mode=poll",
"--repository-poll-secs=5",
"--log-verbose={{ .Values.image.logVerboseLevel }}",
"--allow-gpu-metrics={{ .Values.image.allowGPUMetrics }}"]
env:
- name: AWS_DEFAULT_REGION
valueFrom:
secretKeyRef:
name: aws-credentials
key: AWS_DEFAULT_REGION
- name: AWS_ACCESS_KEY_ID
valueFrom:
secretKeyRef:
name: aws-credentials
key: AWS_ACCESS_KEY_ID
- name: AWS_SECRET_ACCESS_KEY
valueFrom:
secretKeyRef:
name: aws-credentials
key: AWS_SECRET_ACCESS_KEY
- name: AWS_SESSION_TOKEN
valueFrom:
secretKeyRef:
name: aws-credentials
key: AWS_SESSION_TOKEN
ports:
- containerPort: 8000
name: http
- containerPort: 8001
name: grpc
- containerPort: 8002
name: metrics
livenessProbe:
httpGet:
path: /v2/health/live
port: http
initialDelaySeconds: {{ .Values.deployment.livenessProbe.initialDelaySeconds }}
periodSeconds: {{ .Values.deployment.livenessProbe.periodSeconds }}
timeoutSeconds: {{ .Values.deployment.livenessProbe.timeoutSeconds }}
successThreshold: {{ .Values.deployment.livenessProbe.successThreshold }}
failureThreshold: {{ .Values.deployment.livenessProbe.failureThreshold }}
readinessProbe:
httpGet:
path: /v2/health/ready
port: http
initialDelaySeconds: {{ .Values.deployment.readinessProbe.initialDelaySeconds }}
periodSeconds: {{ .Values.deployment.readinessProbe.periodSeconds }}
timeoutSeconds: {{ .Values.deployment.readinessProbe.timeoutSeconds }}
successThreshold: {{ .Values.deployment.readinessProbe.successThreshold }}
failureThreshold: {{ .Values.deployment.readinessProbe.failureThreshold }}
securityContext:
runAsUser: 1000
fsGroup: 1000
| 0 |
rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton | rapidsai_public_repos/cloud-ml-examples/triton/kubernetes/AWS/FIL/helm/charts/triton/templates/istio-vs.yaml | # Copyright (c) 2019-2021, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: triton-vs
spec:
hosts:
- "*"
gateways:
- triton-gateway
tcp:
- match:
- port: 31400
route:
- destination:
host: {{ template "triton-inference-server.name" . }}
port:
number: 8001
http:
- match:
- port: 80
route:
- destination:
host: {{ template "triton-inference-server.name" . }}
port:
number: 8000 | 0 |
rapidsai_public_repos/cloud-ml-examples/optuna | rapidsai_public_repos/cloud-ml-examples/optuna/notebooks/optuna_rapids.ipynb | # ## Run this cell to install optuna and dask_optuna
# Use Optuna 2.3.0 to avoid a bug with DaskStorage (jrbourbeau/dask-optuna#22)
# !pip install optuna==2.3.0 dask_optuna
# ## The plotting libraries
# !pip install plotly
# !pip install -U kaleido
# !pip install 'bokeh<2.0.0'import random
import time
from contextlib import contextmanager
import cudf
import cuml
import dask_cudf
import numpy as np
import optuna
import pandas as pd
import sklearn
import os
import dask
import dask_optuna
from cuml import LogisticRegression
from cuml.model_selection import train_test_split
from cuml.metrics import log_loss
from dask_cuda import LocalCUDACluster
from dask.distributed import Client, wait, performance_report
from joblib import parallel_backend# Helper function for timing blocks of code.
@contextmanager
def timed(name):
t0 = time.time()
yield
t1 = time.time()
print("..%-24s: %8.4f" % (name, t1 - t0))# This will use all GPUs on the local host by default
cluster = LocalCUDACluster(threads_per_worker=1, ip="", dashboard_address="8081")
c = Client(cluster)
# Query the client for all connected workers
workers = c.has_what().keys()
n_workers = len(workers)
cimport os
file_name = 'train.csv'
data_dir = "data/"
INPUT_FILE = os.path.join(data_dir, file_name)N_TRIALS = 150
df = cudf.read_csv(INPUT_FILE)
# Drop ID column
df = df.drop("ID", axis=1)
# Drop non-numerical data and fill NaNs before passing to cuML RF
CAT_COLS = list(df.select_dtypes('object').columns)
df = df.drop(CAT_COLS, axis=1)
df = df.fillna(0)
df = df.astype("float32")
X, y = df.drop(["target"], axis=1), df["target"].astype('int32')
study_name = "dask_optuna_lr_log_loss_tpe"
storage_name = "sqlite:///study_stores.db"def train_and_eval(X_param, y_param, penalty='l2', C=1.0, l1_ratio=None, fit_intercept=True):
"""
Splits the given data into train and test split to train and evaluate the model
for the params parameters.
Params
______
X_param: DataFrame.
The data to use for training and testing.
y_param: Series.
The label for training
penalty, C, l1_ratio, fit_intercept: The parameter values for Logistic Regression.
Returns
score: log loss of the fitted model
"""
X_train, X_valid, y_train, y_valid = train_test_split(X_param,
y_param,
random_state=42)
classifier = LogisticRegression(penalty=penalty,
C=C,
l1_ratio=l1_ratio,
fit_intercept=fit_intercept,
max_iter=10000)
classifier.fit(X_train, y_train)
y_pred = classifier.predict(X_valid)
score = log_loss(y_valid, y_pred)
return scoreprint("Score with default parameters : ",train_and_eval(X, y))def objective(trial, X_param, y_param):
C = trial.suggest_loguniform("C", 0.01, 100.0)
penalty = trial.suggest_categorical("penalty", ['l1', 'none', 'l2'])
fit_intercept = trial.suggest_categorical("fit_intercept", [True, False])
score = train_and_eval(X_param,
y_param,
penalty=penalty,
C=C,
fit_intercept=fit_intercept)
return scorewith timed("dask_optuna"):
# Create a study using Dask-compatible storage
storage = dask_optuna.DaskStorage(storage_name)
study = optuna.create_study(sampler=optuna.samplers.TPESampler(seed=142),
study_name=study_name,
direction="minimize",
storage=storage)
# Optimize in parallel on your Dask cluster
with parallel_backend("dask"):
study.optimize(lambda trial: objective(trial, X, y),
n_trials=N_TRIALS,
n_jobs=n_workers)
print("Number of finished trials: ", len(study.trials))
print("Best trial:")
study.best_trialfrom IPython.display import Imagef = optuna.visualization.plot_param_importances(study)
Image(f.to_image(format="png", engine='kaleido'))f = optuna.visualization.plot_optimization_history(study)
Image(f.to_image(format="png", engine='kaleido'))f = optuna.visualization.plot_parallel_coordinate(study, params=['C', 'penalty', 'fit_intercept'])
Image(f.to_image(format="png", engine='kaleido'))print("Run the following to download the dashboard to dashboard.html: \n\n\t"
f"optuna dashboard --study-name {study_name} --storage '{storage_name}' -o dashboard.html") | 0 |
rapidsai_public_repos/cloud-ml-examples/optuna/notebooks | rapidsai_public_repos/cloud-ml-examples/optuna/notebooks/azure-optuna/run_optuna.ipynb | # verify installation and check Azure ML SDK version
import azureml.core
print('SDK version:', azureml.core.VERSION)from azureml.core.workspace import Workspace
ws = Workspace.from_config()
print('Workspace name: ' + ws.name,
'Azure region: ' + ws.location,
'Subscription id: ' + ws.subscription_id,
'Resource group: ' + ws.resource_group, sep = '\n')
datastore = ws.get_default_datastore()
print("Default datastore's name: {}".format(datastore.name))from azureml.core.compute import ComputeTarget, AmlCompute
from azureml.core.compute_target import ComputeTargetException
# choose a name for your cluster
gpu_cluster_name = 'gpu-cluster'
if gpu_cluster_name in ws.compute_targets:
gpu_cluster = ws.compute_targets[gpu_cluster_name]
if gpu_cluster and type(gpu_cluster) is AmlCompute:
print('Found compute target. Will use {0} '.format(gpu_cluster_name))
else:
print('creating new cluster')
# m_size parameter below could be modified to one of the RAPIDS-supported VM types
provisioning_config = AmlCompute.provisioning_configuration(vm_size = 'Standard_NC6s_v3', max_nodes = 5, idle_seconds_before_scaledown = 300)
# Use VM types with more than one GPU for multi-GPU option, e.g. Standard_NC12s_v3
# create the cluster
gpu_cluster = ComputeTarget.create(ws, gpu_cluster_name, provisioning_config)
# can poll for a minimum number of nodes and for a specific timeout
# if no min node count is provided it uses the scale settings for the cluster
gpu_cluster.wait_for_completion(show_output=True, min_node_count=None, timeout_in_minutes=20)
# use get_status() to get a detailed status for the current cluster
print(gpu_cluster.get_status().serialize())from azureml.core import Experiment
experiment_name = 'optuna_rapids'
experiment = Experiment(ws, name=experiment_name)from azureml.core import Environment
# create the environment
rapids_env = Environment('rapids_env')
# create the environment inside a Docker container
rapids_env.docker.enabled = True
# specify docker steps as a string. Alternatively, load the string from a file
dockerfile = """
FROM rapidsai/rapidsai:21.06-cuda10.2-runtime-ubuntu18.04-py3.7
RUN apt-get update && \
apt-get install -y fuse && \
apt-get install libssl1.0.0 libssl-dev && \
source activate rapids && \
pip install azureml-sdk==1.13.0 && \
pip install azureml-widgets && \
pip install optuna && \
pip install dask_optuna && \
pip install fusepy
"""
# set base image to None since the image is defined by dockerfile
rapids_env.docker.enabled = True
rapids_env.docker.base_image = None
rapids_env.docker.base_dockerfile = dockerfile
# use rapids environment in the container
rapids_env.python.user_managed_dependencies = Truefrom azureml.core.dataset import Dataset
data_dir = "data/"
path_on_datastore = 'bnp_upload'
datastore.upload(src_dir=data_dir, target_path=path_on_datastore, overwrite=False, show_progress=True)
ds_data = datastore.path(path_on_datastore)
dataset = Dataset.File.from_files(ds_data)script_params = ['--data_dir', dataset.as_named_input('bnp_input').as_mount(),
]
from azureml.core import ScriptRunConfig
project_folder = "project_folder/"
src = ScriptRunConfig(source_directory=project_folder,
script='train_optuna.py',
arguments=script_params,
compute_target="gpu-cluster",
environment=rapids_env)
run = experiment.submit(config=src)
from azureml.widgets import RunDetails
RunDetails(run).show()
run.wait_for_completion(show_output=False) | 0 |
rapidsai_public_repos/cloud-ml-examples/optuna/notebooks/azure-optuna | rapidsai_public_repos/cloud-ml-examples/optuna/notebooks/azure-optuna/project_folder/train_optuna.py | import random
import time
from contextlib import contextmanager
import cudf
import cuml
import dask_cudf
import numpy as np
import optuna
import pandas as pd
import sklearn
import os
import dask
import dask_optuna
from cuml import LogisticRegression
from cuml.model_selection import train_test_split
from cuml.metrics import log_loss
from dask_cuda import LocalCUDACluster
from dask.distributed import Client, wait, performance_report
from joblib import parallel_backend
import argparse
from azureml.core.run import Run
run = Run.get_context()
def train_and_eval(X_param, y_param, penalty='l2', C=1.0, l1_ratio=None, fit_intercept=True):
"""
Splits the given data into train and test split to train and evaluate the model
for the params parameters.
Params
______
X_param: DataFrame.
The data to use for training and testing.
y_param: Series.
The label for training
penalty, C, l1_ratio, fit_intercept: The parameter values for Logistic Regression.
Returns
score: log loss of the fitted model
"""
X_train, X_valid, y_train, y_valid = train_test_split(X_param,
y_param,
random_state=42)
classifier = LogisticRegression(penalty=penalty,
C=C,
l1_ratio=l1_ratio,
fit_intercept=fit_intercept)
classifier.fit(X_train, y_train)
y_pred = classifier.predict(X_valid)
score = log_loss(y_valid, y_pred)
return score
def objective(trial, X_param, y_param):
"""
Objective function used for Optuna experiments.
Returns the score for optimization task.
"""
C = trial.suggest_uniform("C", 0 , 9.0)
penalty = trial.suggest_categorical("penalty", ['l1', 'none', 'l2'])
l1_ratio = trial.suggest_uniform("l1_ratio", 0 , 1.0)
fit_intercept = trial.suggest_categorical("fit_intercept", [True, False])
score = train_and_eval(X_param,
y_param,
penalty=penalty,
C=C,
l1_ratio=l1_ratio,
fit_intercept=fit_intercept)
run.log('C', np.float(C))
run.log('penalty', penalty)
run.log('l1_ratio', np.float(l1_ratio))
run.log('fit_intercept', fit_intercept)
run.log('score',np.float(score))
return score
if __name__=="__main__":
parser = argparse.ArgumentParser()
parser.add_argument('--data_dir', type=str, help='location of data')
args = parser.parse_args()
data_dir = args.data_dir
print('Data folder is at:', data_dir)
print('List all files: ', os.listdir(data_dir))
cluster = LocalCUDACluster(threads_per_worker=1, ip="", dashboard_address="8088")
c = Client(cluster)
# Query the client for all connected workers
workers = c.has_what().keys()
n_workers = len(workers)
df = cudf.read_csv(os.path.join(data_dir, "train.csv"))
N_TRIALS = 5
# Drop non-numerical data and fill NaNs before passing to cuML RF
CAT_COLS = list(df.select_dtypes('object').columns)
df = df.drop(CAT_COLS, axis=1)
df = df.dropna()
df = df.astype("float32")
X, y = df.drop(["target"], axis=1), df["target"].astype('int32')
study_name = "dask_optuna_lr_log_loss_tpe"
storage_name = "sqlite:///study_stores.db"
storage = dask_optuna.DaskStorage(storage_name)
study = optuna.create_study(sampler=optuna.samplers.TPESampler(),
study_name=study_name,
direction="minimize",
storage=storage)
# Optimize in parallel on your Dask cluster
with parallel_backend("dask"):
study.optimize(lambda trial: objective(trial, X, y),
n_trials=N_TRIALS,
n_jobs=n_workers)
print('Best params{} and best score{}'.format(study.best_params, study.best_value)) | 0 |
rapidsai_public_repos/cloud-ml-examples/optuna/notebooks | rapidsai_public_repos/cloud-ml-examples/optuna/notebooks/jupytercon_demo/RAPIDS_xfeat_Optuna-CPU.ipynb | import time
import json
import requests
import logging
import numpy as np
import mlflow
import mlflow.sklearn
from mlflow.tracking import MlflowClient
from mlflow.models.signature import infer_signature
import optuna
from optuna.integration.mlflow import MLflowCallback
from optuna.study import StudyDirection
from optuna.trial import TrialState
from optuna import type_checking
import xfeat
from xfeat.pipeline import Pipeline
from xfeat.num_encoder import SelectNumerical
from xfeat.selector import ChiSquareKBest
from xfeat.optuna_selector import KBestThresholdExplorer
from functools import partial
from sklearn.metrics import roc_auc_score
from sklearn.linear_model import LogisticRegression
import sklearn
import pandas as pd
from sklearn.model_selection import train_test_split
from xfeat.cat_encoder import LabelEncoder, TargetEncoder
from xfeat import ArithmeticCombinations, Pipeline, SelectNumericalimport time
from contextlib import contextmanager
# Helping time blocks of code
@contextmanager
def timed(txt):
t0 = time.time()
yield
t1 = time.time()
print("%32s time: %8.5f" % (txt, t1 - t0))MLFLOW_TRACKING_URI='sqlite:////tmp/mlflow-db.sqlite'
MLFLOW_MODEL_ID = "rapids-optuna-airline"
def get_latest_mlflow_model(tracking_uri, model_id):
client = MlflowClient(tracking_uri=tracking_uri, registry_uri=tracking_uri)
model = client.get_registered_model(model_id)
latest_model = model.latest_versions[0]
return f"MLFLOW_TRACKING_URI={tracking_uri} mlflow models serve --no-conda -m models:/{model_id}/{latest_model.version} -p 56767"
## Custom callback, for additional flexibility, based on MLflowCallback
class RAPIDSMLflowCallback(object):
def __init__(self, tracking_uri: str = "sqlite:////tmp/mlflow-db.sqlite",
experiment_name: str = "RAPIDS-Optuna",
metric_name="value"):
self._tracking_uri = tracking_uri
self._experiment_name = experiment_name
self._metric_name = metric_name
def __call__(self, study, trial):
if (self._tracking_uri is not None):
mlflow.set_tracking_uri(self._tracking_uri)
eid = mlflow.set_experiment(self._experiment_name)
with mlflow.start_run(run_name=f"Trial: {trial.number}", experiment_id=eid, nested=True):
trial_value = trial.value if trial.value is not None else float("nan")
mlflow.log_metric(self._metric_name, trial_value)
mlflow.log_params(trial.params)
tags = {}
tags["number"] = str(trial.number)
tags["datetime_start"] = str(trial.datetime_start)
tags["datetime_complete"] = str(trial.datetime_complete)
tags['SKlearn Version'] = str(sklearn.__version__)
trial_state = trial.state
if (isinstance(trial_state, TrialState)):
tags['state'] = str(trial_state).split('.')[-1]
# Set direction and convert it to str and remove the common prefix.
study_direction = study.direction
if isinstance(study_direction, StudyDirection):
tags["direction"] = str(study_direction).split(".")[-1]
tags.update(trial.user_attrs)
distributions = {
(k + "_distribution"): str(v) for (k, v) in trial.distributions.items()
}
tags.update(distributions)
# This is a temporary fix on Optuna side. It avoids an error with user
# attributes that are too long. It should be fixed on MLflow side later.
# When it is fixed on MLflow side this codeblock can be removed.
# see https://github.com/optuna/optuna/issues/1340
# see https://github.com/mlflow/mlflow/issues/2931
max_mlflow_tag_length = 5000
for key, value in tags.items():
value = str(value) # make sure it is a string
if len(value) > max_mlflow_tag_length:
tags[key] = textwrap.shorten(value, max_mlflow_tag_length)
mlflow.set_tags(tags) def feature_engineering(df_train, df_test):
"""
Perform feature engineering and return a new df with engineered features
"""
encoder = Pipeline([
LabelEncoder(output_suffix=""),
TargetEncoder(target_col=TARGET_COL, output_suffix=""),
SelectNumerical(),
ArithmeticCombinations(exclude_cols=[TARGET_COL],
drop_origin=False,
operator="+",
r=2,
output_suffix="_plus")
])
df_train = encoder.fit_transform(df_train)
df_test = encoder.transform(df_test)
return df_train, df_testdef train_and_eval(df_train, df_test, penalty='l2', C=1.0, l1_ratio=None, fit_intercept=True, selector=None, return_model=False):
if selector:
# Getting the label column as it is dropped in the selector
y_train = df_train[TARGET_COL]
y_test = df_test[TARGET_COL]
X_train = selector.fit_transform(df_train)
X_test = selector.transform(df_test)
else:
X_train = df_train[df_train.columns.difference([TARGET_COL])]
X_test = df_test[df_test.columns.difference([TARGET_COL])]
y_train = df_train[TARGET_COL]
y_test = df_test[TARGET_COL]
classifier = LogisticRegression(solver='saga',
penalty=penalty,
C=C,
l1_ratio=l1_ratio,
fit_intercept=fit_intercept)
classifier.fit(X_train, y_train)
y_pred = classifier.predict(X_test)
score = roc_auc_score(y_test, y_pred)
if (return_model):
return score, classifier, infer_signature(X_test, y_pred)
return score
def objective(df_train, df_test, selector, trial):
"""
Performs the training and evaluation of the set of parameters and subset of features using selector.
"""
selector.set_trial(trial)
# Select Params
C = trial.suggest_uniform("C", 0 , 9.0)
penalty = trial.suggest_categorical("penalty", ['l1', 'none', 'l2'])
l1_ratio = trial.suggest_uniform("l1_ratio", 0 , 1.0)
fit_intercept = trial.suggest_categorical("fit_intercept", [True, False])
score = train_and_eval(df_train,
df_test,
penalty=penalty,
C=C,
l1_ratio=l1_ratio,
fit_intercept=fit_intercept,
selector=selector)
return score
def feature_selection(df_train, df_test, experiment_name):
"""
Defines the Pipeline and performs the optuna opt
"""
artifact_path = "rapids-optuna-airline"
selector = Pipeline(
[
SelectNumerical(),
KBestThresholdExplorer(ChiSquareKBest(target_col=TARGET_COL)),
]
)
mlfcb = RAPIDSMLflowCallback(
tracking_uri=MLFLOW_TRACKING_URI,
experiment_name=experiment_name,
metric_name='auc')
study = optuna.create_study(direction="maximize")
mlflow.set_tracking_uri(MLFLOW_TRACKING_URI)
mlflow.set_experiment(experiment_name)
with mlflow.start_run(run_name=f"Optuna-HPO:{study.study_name}"):
study.optimize(partial(objective, df_train, df_test, selector), n_trials=N_TRIALS, callbacks=[mlfcb])
selector.from_trial(study.best_trial)
selected_cols = selector.get_selected_cols()
df_select_train = df_train[selected_cols]
df_select_train[TARGET_COL] = df_train[TARGET_COL]
df_select_test = df_test[selected_cols]
df_select_test[TARGET_COL] = df_test[TARGET_COL]
params = study.best_params
score, classifier, signature = train_and_eval(df_select_train,
df_select_test,
C=params['C'],
penalty=params['penalty'],
l1_ratio=params['l1_ratio'],
fit_intercept=params['fit_intercept'],
return_model=True)
with mlflow.start_run(run_name='Final Classifier', nested=True):
mlflow.log_metric('auc', score)
mlflow.log_params(params)
mlflow.sklearn.log_model(classifier,
signature=signature,
artifact_path=artifact_path,
registered_model_name="rapids-optuna-airline",
conda_env='conda/conda.yaml')
df_select = pd.concat([df_select_train, df_select_test], sort=False)
return study, df_select.reset_index(drop=True), classifier, scoreINPUT_FILE = "data/airline_small.parquet" # Change to correspond to local path
N_ROWS = 100000
N_TRIALS = 10
TARGET_COL = "ArrDelayBinary"start_time = time.time()
df_pandas = pd.read_parquet(INPUT_FILE)[:N_ROWS]
df_train, df_test = train_test_split(df_pandas,
random_state=np.random.seed(74),
shuffle=True)
print("Default scikit-learn ", train_and_eval(df_train, df_test))
with timed("Pandas FE"):
df_feature_eng_train, df_feature_eng_test = feature_engineering(df_train, df_test)
df_feature_eng_train[TARGET_COL] = df_feature_eng_train[TARGET_COL].astype('float64')
df_feature_eng_test[TARGET_COL] = df_feature_eng_test[TARGET_COL].astype('float64')
score = train_and_eval(df_feature_eng_train, df_feature_eng_test)
print("After feature eng: ", score)
with timed("FS + Optuna PANDAS"):
# Disable Alembic driver, used by MLflow, from logging INFO messages to the command line.
logging.getLogger('alembic').setLevel(logging.CRITICAL)
study, df_select, best_clf, score = feature_selection(df_feature_eng_train,
df_feature_eng_test,
experiment_name="Pandas")
print("Best score after Feature Selection + Optuna: ", score)
print("Complete Pandas ", time.time() - start_time)print("The details of the best trial ", study.best_trial)from IPython.display import Image
f = optuna.visualization.plot_param_importances(study)
Image(f.to_image(format="png", engine='kaleido'))f = optuna.visualization.plot_slice(study, params=['l1_ratio', 'C', 'KBestThresholdExplorer.k', 'penalty', 'fit_intercept'])
Image(f.to_image(format="png", engine='kaleido'))f = optuna.visualization.plot_optimization_history(study)
Image(f.to_image(format="png", engine='kaleido')) | 0 |
rapidsai_public_repos/cloud-ml-examples/optuna/notebooks | rapidsai_public_repos/cloud-ml-examples/optuna/notebooks/jupytercon_demo/RAPIDS_xfeat_Optuna-GPU.ipynb | # !pip install mlflow
# !pip install optuna
# !pip install plotly
# !pip install kaleido
# !pip install xfeatimport time
import json
import requests
import logging
from contextlib import contextmanager
import numpy as np
import cupy
import cudf
import cuml
from cuml import LogisticRegression
from cuml.metrics import roc_auc_score
from cuml.model_selection import train_test_split
from cuml.preprocessing.LabelEncoder import LabelEncoder
from cuml.preprocessing.TargetEncoder import TargetEncoder
import mlflow
import mlflow.sklearn
from mlflow.tracking import MlflowClient
from mlflow.models.signature import infer_signature
import optuna
import sklearn
from optuna.integration.mlflow import MLflowCallback
from optuna.study import StudyDirection
from optuna.trial import TrialState
from optuna import type_checking
import xfeat
from xfeat.pipeline import Pipeline
from xfeat.num_encoder import SelectNumerical
from xfeat.selector import ChiSquareKBest
from xfeat.optuna_selector import KBestThresholdExplorer
from xfeat import ArithmeticCombinations, Pipeline, SelectNumerical
from functools import partial# Helping time blocks of code
@contextmanager
def timed(txt):
t0 = time.time()
yield
t1 = time.time()
print("%32s time: %8.5f" % (txt, t1 - t0))import os
download_data = True
file_name = 'airline_small.parquet' # NOTE: Change to airline_20000000.parquet to use a larger dataset
data_dir = "data/" # NOTE: Change to a local path where you want to save the file
INPUT_FILE = os.path.join(data_dir, file_name)if download_data:
from urllib.request import urlretrieve
if os.path.isfile(INPUT_FILE):
print(f" > File already exists. Ready to load at {INPUT_FILE}")
else:
# Ensure folder exists
os.makedirs(data_dir, exist_ok=True)
url = "https://rapidsai-cloud-ml-sample-data.s3-us-west-2.amazonaws.com/" + file_name
urlretrieve(url= url,filename=INPUT_FILE)
print("Completed!")MLFLOW_TRACKING_URI='sqlite:////tmp/mlflow-db.sqlite'
MLFLOW_MODEL_ID = "rapids-optuna-airline"
def get_latest_mlflow_model(tracking_uri, model_id):
client = MlflowClient(tracking_uri=tracking_uri, registry_uri=tracking_uri)
model = client.get_registered_model(model_id)
latest_model = model.latest_versions[0]
return f"MLFLOW_TRACKING_URI={tracking_uri} mlflow models serve --no-conda -m models:/{model_id}/{latest_model.version} -p 56767"
## Custom callback, for additional flexibility, based on MLflowCallback
class RAPIDSMLflowCallback(object):
def __init__(self, tracking_uri: str = "sqlite:////tmp/mlflow-db.sqlite",
experiment_name: str = "RAPIDS-Optuna",
metric_name="value"):
self._tracking_uri = tracking_uri
self._experiment_name = experiment_name
self._metric_name = metric_name
def __call__(self, study, trial):
if (self._tracking_uri is not None):
mlflow.set_tracking_uri(self._tracking_uri)
eid = mlflow.set_experiment(self._experiment_name)
with mlflow.start_run(run_name=f"Trial: {trial.number}", experiment_id=eid, nested=True):
trial_value = trial.value if trial.value is not None else float("nan")
mlflow.log_metric(self._metric_name, trial_value)
mlflow.log_params(trial.params)
tags = {}
tags["number"] = str(trial.number)
tags["datetime_start"] = str(trial.datetime_start)
tags["datetime_complete"] = str(trial.datetime_complete)
tags['RAPIDS cuDF Version'] = str(cudf.__version__)
tags['RAPIDS cuML Version'] = str(cuml.__version__)
tags['SKlearn Version'] = str(sklearn.__version__)
trial_state = trial.state
if (isinstance(trial_state, TrialState)):
tags['state'] = str(trial_state).split('.')[-1]
# Set direction and convert it to str and remove the common prefix.
study_direction = study.direction
if isinstance(study_direction, StudyDirection):
tags["direction"] = str(study_direction).split(".")[-1]
tags.update(trial.user_attrs)
distributions = {
(k + "_distribution"): str(v) for (k, v) in trial.distributions.items()
}
tags.update(distributions)
# This is a temporary fix on Optuna side. It avoids an error with user
# attributes that are too long. It should be fixed on MLflow side later.
# When it is fixed on MLflow side this codeblock can be removed.
# see https://github.com/optuna/optuna/issues/1340
# see https://github.com/mlflow/mlflow/issues/2931
max_mlflow_tag_length = 5000
for key, value in tags.items():
value = str(value) # make sure it is a string
if len(value) > max_mlflow_tag_length:
tags[key] = textwrap.shorten(value, max_mlflow_tag_length)
mlflow.set_tags(tags) def feature_engineering(df_train, df_test):
"""
Perform feature engineering and return a new train and test df with engineered features
"""
for col in CAT_COLS:
out_col = f'{col}_TE'
lbl_enc = LabelEncoder(handle_unknown='ignore')
tar_enc = TargetEncoder(n_folds=5, smooth=TARGET_ENC_SMOOTH, split_method=TARGET_ENC_SPLIT)
df_train[col] = lbl_enc.fit_transform(df_train[col])
df_test[col] = lbl_enc.transform(df_test[col])
df_train[col] = df_train[col].fillna(0)
df_test[col] = df_test[col].fillna(0)
df_train[out_col] = tar_enc.fit_transform(df_train[col], df_train[TARGET_COL])
df_test[out_col] = tar_enc.transform(df_test[col])
del lbl_enc, tar_enc
encoder = Pipeline([
SelectNumerical(),
ArithmeticCombinations(exclude_cols=[TARGET_COL],
drop_origin=False,
operator="+",
r=2,
output_suffix="_plus")
])
df_train = encoder.fit_transform(df_train)
df_test = encoder.transform(df_test)
return df_train, df_testdef train_and_eval(df_train, df_test, penalty='l2', C=1.0, l1_ratio=None, fit_intercept=True, selector=None, return_model=False):
"""
Split the dataframe based on TARGET_COL
Accepts the parameters to be set for the Logistic Regression model
Evaluates on the split data and returns AUC score.
Parameters
----------
df_train, df_test: DataFrame to use for training and validation
penalty, C, l1_ratio, fit_intercept: Parameters for the LogisticRegression Model
For details refer to the documentation
selector: xfeat selector passed via Optuna (default=None)
When set, the slector is used to fit and transform the data
return_model: Returns the fit model back to the calling fucntion (default=False)
Returns
----------
score: AUC score of the validation set
classifier, signature: Returned only when return_model is set to True
"""
if selector:
# Getting the label column as it is dropped in the selector
y_train = df_train[TARGET_COL]
y_test = df_test[TARGET_COL]
X_train = selector.fit_transform(df_train)
X_test = selector.transform(df_test)
else:
y_train = df_train[TARGET_COL]
y_test = df_test[TARGET_COL]
X_train = df_train[df_train.columns.difference([TARGET_COL])]
X_test = df_test[df_test.columns.difference([TARGET_COL])]
# Train and get accuracy
classifier = LogisticRegression(penalty=penalty,
C=C,
l1_ratio=l1_ratio,
fit_intercept=fit_intercept)
classifier.fit(X_train, y_train)
y_pred = classifier.predict_proba(X_test.values)[:, 1]
y_pred = y_pred.astype(y_test.dtype)
score = roc_auc_score(y_test, y_pred)
if (return_model):
return score, classifier, infer_signature(X_test.to_pandas(), cupy.asnumpy(y_pred))
return score
def objective(df_train, df_test, selector, trial):
"""
Performs the training and evaluation of the set of parameters and subset of features using selector.
"""
selector.set_trial(trial)
# Select Params
C = trial.suggest_uniform("C", 0 , 9.0)
penalty = trial.suggest_categorical("penalty", ['l1', 'none', 'l2'])
l1_ratio = trial.suggest_uniform("l1_ratio", 0 , 1.0)
fit_intercept = trial.suggest_categorical("fit_intercept", [True, False])
score = train_and_eval(df_train,
df_test,
penalty=penalty,
C=C,
l1_ratio=l1_ratio,
fit_intercept=fit_intercept,
selector=selector)
return score
def feature_selection(df_train, df_test, experiment_name):
"""
Defines the Pipeline and performs the optuna opt
"""
artifact_path = "rapids-optuna-airline"
selector = Pipeline(
[
SelectNumerical(),
# Select features according to the k highest scores from the ChiSquared test
KBestThresholdExplorer(ChiSquareKBest(target_col=TARGET_COL)),
]
)
mlfcb = RAPIDSMLflowCallback(
tracking_uri=MLFLOW_TRACKING_URI,
experiment_name=experiment_name,
metric_name='auc')
study = optuna.create_study(direction="maximize")
mlflow.set_tracking_uri(MLFLOW_TRACKING_URI)
mlflow.set_experiment(experiment_name)
with mlflow.start_run(run_name=f"Optuna-HPO:{study.study_name}"):
study.optimize(partial(objective, df_train, df_test, selector), n_trials=N_TRIALS, callbacks=[mlfcb])
selector.from_trial(study.best_trial)
selected_cols = selector.get_selected_cols()
df_select_train = df_train[selected_cols]
df_select_train[TARGET_COL] = df_train[TARGET_COL]
df_select_test = df_test[selected_cols]
df_select_test[TARGET_COL] = df_test[TARGET_COL]
params = study.best_params
score, classifier, signature = train_and_eval(df_select_train,
df_select_test,
C=params['C'],
penalty=params['penalty'],
l1_ratio=params['l1_ratio'],
fit_intercept=params['fit_intercept'],
return_model=True)
with mlflow.start_run(run_name='Final Classifier', nested=True):
mlflow.log_metric('auc', score)
mlflow.log_params(params)
mlflow.sklearn.log_model(classifier,
signature=signature,
artifact_path=artifact_path,
registered_model_name="rapids-optuna-airline",
conda_env='conda/conda.yaml')
df_select = cudf.concat([df_select_train, df_select_test], sort=False)
return study, df_select.reset_index(drop=True), classifier, scoreN_ROWS = 100000 # Number of rows to use for this experiment run
N_TRIALS = 50 # Number of trials for the HPO study
TARGET_COL = "ArrDelayBinary"
CAT_COLS = ["Dest", "Origin", "UniqueCarrier"]
# Parameters for TagetEncoder
TARGET_ENC_SMOOTH = 0.001
TARGET_ENC_SPLIT = 'interleaved'start_time = time.time()
df_ = cudf.read_parquet(INPUT_FILE)[:N_ROWS]
# Can't handle nagative values, yet
df_ = df_.drop(["ActualElapsedTime"], axis=1)
df_train, df_test = train_test_split(df_,
random_state=np.random.seed(0),
shuffle=True)
print("Default performance: ", train_and_eval(df_train, df_test))
df_.head()# We cast to objects for categorical and target encoding
# Can't pass categorical directly to LR
for col in CAT_COLS:
df_[col] = df_[col].astype("object")
with timed("Feature Engineering"):
df_feature_eng_train, df_feature_eng_test = feature_engineering(df_train, df_test)
df_feature_eng_train[TARGET_COL] = df_feature_eng_train[TARGET_COL].astype('float32')
df_feature_eng_test[TARGET_COL] = df_feature_eng_test[TARGET_COL].astype('float32')
score = train_and_eval(df_feature_eng_train, df_feature_eng_test)
print("After feature eng: ", score)
df_feature_eng_train.head()import random
exp_name = 'RAPIDS-Optuna-Single-GPU' + str(random.randint(0,100))
with timed("Feature Selection + Optuna"):
# Disable Alembic driver, used by MLflow, from logging INFO messages to the command line.
logging.getLogger('alembic').setLevel(logging.CRITICAL)
study, df_select, best_clf, score = feature_selection(df_feature_eng_train,
df_feature_eng_test,
experiment_name=exp_name)
print("Best score after Feature Selection + Optuna: ", score)
df_select.head()end_time = time.time()
print("Complete workflow ", end_time - start_time)print("The details of the best trial ", study.best_trial)from IPython.display import Image
f = optuna.visualization.plot_param_importances(study)
Image(f.to_image(format="png", engine='kaleido'))f = optuna.visualization.plot_slice(study, params=['l1_ratio', 'C', 'KBestThresholdExplorer.k', 'penalty', 'fit_intercept'])
Image(f.to_image(format="png", engine='kaleido'))f = optuna.visualization.plot_optimization_history(study)
Image(f.to_image(format="png", engine='kaleido'))print(f"Run the command below in a terminal, and wait for it to load your model:\n\n \
{get_latest_mlflow_model(MLFLOW_TRACKING_URI, MLFLOW_MODEL_ID)}")host='localhost'
port='56767'
headers = {
"Content-Type": "application/json",
"format": "pandas-split"
}
data = {
"columns": ["Year", "Month", "DayofMonth", "DayofWeek", "CRSDepTime", "CRSArrTime", "UniqueCarrier",
"FlightNum", "ActualElapsedTime", "Origin", "Dest", "Distance", "Diverted"],
"data": [[1987, 10, 1, 4, 1, 556, 0, 190, 247, 202, 162, 1846, 0]]
}
while (True):
try:
resp = requests.post(url="http://%s:%s/invocations" % (host, port), data=json.dumps(data), headers=headers)
print('Classification: %s' % ("ON-Time" if resp.text == "[0.0]" else "LATE"))
break
except Exception as e:
errmsg = "Caught exception attempting to call model endpoint: %s" % e
print(errmsg)
print("... Sleeping ...")
time.sleep(20) | 0 |
rapidsai_public_repos/cloud-ml-examples/optuna/notebooks/jupytercon_demo | rapidsai_public_repos/cloud-ml-examples/optuna/notebooks/jupytercon_demo/conda/conda.yaml | name: xfeatOptuna
channels:
- rapidsai-nightly
- nvidia
- conda-forge
dependencies:
- _libgcc_mutex=0.1=conda_forge
- _openmp_mutex=4.5=0_gnu
- abseil-cpp=20200225.2=he1b5a44_1
- aiohttp=3.6.2=py37h516909a_0
- appdirs=1.4.3=py_1
- arrow-cpp=0.17.1=py37h1234567_11_cuda
- arrow-cpp-proc=1.0.0=cuda
- async-timeout=3.0.1=py_1000
- attrs=19.3.0=py_0
- aws-sdk-cpp=1.7.164=hc831370_1
- backcall=0.2.0=pyh9f0ad1d_0
- bleach=3.1.5=pyh9f0ad1d_0
- blosc=1.19.0=he1b5a44_0
- bokeh=2.1.1=py37hc8dfbb8_0
- boost=1.72.0=py37h9de70de_0
- boost-cpp=1.72.0=h8e57a91_0
- brotli=1.0.7=he1b5a44_1004
- brotlipy=0.7.0=py37h8f50634_1000
- brunsli=0.1=he1b5a44_0
- bzip2=1.0.8=h516909a_2
- c-ares=1.16.1=h516909a_0
- ca-certificates=2020.6.20=hecda079_0
- cairo=1.16.0=hcf35c78_1003
- certifi=2020.6.20=py37hc8dfbb8_0
- cffi=1.14.0=py37he30daa8_1
- cfitsio=3.470=h3eac812_5
- chardet=3.0.4=py37hc8dfbb8_1006
- charls=2.1.0=he1b5a44_2
- click=7.1.2=pyh9f0ad1d_0
- click-plugins=1.1.1=py_0
- cligj=0.5.0=py_0
- cloudpickle=1.5.0=py_0
- colorcet=2.0.1=py_0
- cryptography=3.0=py37hb09aad4_0
- cudatoolkit=10.2.89=h6bb024c_0
- cudf=0.15.0a200721=py37_g79bd43517_3063
- cudnn=7.6.5=cuda10.2_0
- cugraph=0.15.0a200721=py37_g7c1db5c0_560
- cuml=0.15.0a200721=cuda10.2_py37_g8c602c05f_1268
- cupy=7.6.0=py37h940342b_0
- curl=7.71.1=he644dc0_3
- cusignal=0.15.0a200721=py37_g12d5d2e_299
- cuspatial=0.15.0a200721=py37_g688a604_257
- cuxfilter=0.15.0a200721=py37_g1d5c035_171
- cycler=0.10.0=py_2
- cytoolz=0.10.1=py37h516909a_0
- dask=2.21.0=py_0
- dask-core=2.21.0=py_0
- dask-cuda=0.15.0a200721=py37_76
- dask-cudf=0.15.0a200721=py37_g79bd43517_3063
- dask-xgboost=0.2.0.dev28=cuda10.2py37_0
- datashader=0.10.0=py_0
- datashape=0.5.4=py_1
- decorator=4.4.2=py_0
- defusedxml=0.6.0=py_0
- distributed=2.21.0=py37hc8dfbb8_0
- dlpack=0.3=he1b5a44_1
- double-conversion=3.1.5=he1b5a44_2
- entrypoints=0.3=py37hc8dfbb8_1001
- expat=2.2.9=he1b5a44_2
- fastavro=0.23.6=py37h8f50634_0
- fastrlock=0.5=py37h3340039_0
- fiona=1.8.13=py37h0492a4a_1
- fontconfig=2.13.1=h86ecdb6_1001
- freetype=2.10.2=he06d7ca_0
- freexl=1.0.5=h516909a_1002
- fsspec=0.7.4=py_0
- gdal=3.0.4=py37h4b180d9_10
- geopandas=0.8.1=py_0
- geos=3.8.1=he1b5a44_0
- geotiff=1.6.0=h05acad5_0
- gflags=2.2.2=he1b5a44_1004
- giflib=5.2.1=h516909a_2
- glib=2.65.0=h3eb4bd4_0
- glog=0.4.0=h49b9bf7_3
- grpc-cpp=1.30.1=h9ea6770_0
- hdf4=4.2.13=hf30be14_1003
- hdf5=1.10.6=nompi_h3c11f04_100
- heapdict=1.0.1=py_0
- icu=64.2=he1b5a44_1
- idna=2.10=pyh9f0ad1d_0
- imagecodecs=2020.5.30=py37hda6ee5b_2
- imageio=2.9.0=py_0
- importlib-metadata=1.7.0=py37hc8dfbb8_0
- importlib_metadata=1.7.0=0
- ipykernel=5.3.3=py37h43977f1_0
- ipython=7.16.1=py37h43977f1_0
- ipython_genutils=0.2.0=py_1
- ipywidgets=7.5.1=py_0
- jedi=0.17.2=py37hc8dfbb8_0
- jinja2=2.11.2=pyh9f0ad1d_0
- joblib=0.16.0=py_0
- jpeg=9d=h516909a_0
- json-c=0.13.1=hbfbb72e_1002
- jsonschema=3.2.0=py37hc8dfbb8_1
- jupyter-server-proxy=1.5.0=py_0
- jupyter_client=6.1.6=py_0
- jupyter_core=4.6.3=py37hc8dfbb8_1
- jxrlib=1.1=h516909a_2
- kealib=1.4.13=h33137a7_1
- kiwisolver=1.2.0=py37h99015e2_0
- krb5=1.17.1=hfafb76e_1
- lcms2=2.11=hbd6801e_0
- ld_impl_linux-64=2.33.1=h53a641e_7
- lerc=2.2=he1b5a44_0
- libaec=1.0.4=he1b5a44_1
- libblas=3.8.0=17_openblas
- libcblas=3.8.0=17_openblas
- libcudf=0.15.0a200721=cuda10.2_g79bd43517_3063
- libcugraph=0.15.0a200721=cuda10.2_g7c1db5c0_560
- libcuml=0.15.0a200721=cuda10.2_g8c602c05f_1268
- libcumlprims=0.15.0a200622=cuda10.2_45
- libcurl=7.71.1=hcdd3856_3
- libcuspatial=0.15.0a200721=cuda10.2_g688a604_257
- libdap4=3.20.6=h1d1bd15_1
- libedit=3.1.20191231=h14c3975_1
- libevent=2.1.10=hcdb4288_1
- libffi=3.3=he6710b0_2
- libgcc-ng=9.2.0=h24d8f2e_2
- libgdal=3.0.4=he6a97d6_10
- libgfortran-ng=7.5.0=hdf63c60_6
- libgomp=9.2.0=h24d8f2e_2
- libhwloc=2.1.0=h3c4fd83_0
- libiconv=1.15=h516909a_1006
- libkml=1.3.0=hb574062_1011
- liblapack=3.8.0=17_openblas
- libllvm9=9.0.1=he513fc3_1
- libnetcdf=4.7.4=nompi_h84807e1_104
- libopenblas=0.3.10=pthreads_hb3c22a3_3
- libpng=1.6.37=hed695b0_1
- libpq=12.3=h5513abc_0
- libprotobuf=3.12.3=h8b12597_2
- librmm=0.15.0a200721=cuda10.2_ge2c0e95_368
- libsodium=1.0.17=h516909a_0
- libspatialindex=1.9.3=he1b5a44_3
- libspatialite=4.3.0a=h2482549_1038
- libssh2=1.9.0=hab1572f_4
- libstdcxx-ng=9.1.0=hdf63c60_0
- libtiff=4.1.0=hc7e4089_6
- libuuid=2.32.1=h14c3975_1000
- libuv=1.34.0=h516909a_0
- libwebp=1.1.0=h56121f0_4
- libwebp-base=1.1.0=h516909a_3
- libxcb=1.13=h14c3975_1002
- libxgboost=1.1.0dev.rapidsai0.15=cuda10.2_1
- libxml2=2.9.10=hee79883_0
- libzopfli=1.0.3=he1b5a44_0
- llvm-tools=9.0.1=he513fc3_1
- llvmdev=9.0.1=he513fc3_1
- llvmlite=0.33.0=py37h5202443_0
- locket=0.2.0=py_2
- lz4-c=1.9.2=he1b5a44_1
- markdown=3.2.2=py_0
- markupsafe=1.1.1=py37h8f50634_1
- matplotlib-base=3.3.0=py37hbe0a388_0
- mistune=0.8.4=py37h8f50634_1001
- msgpack-python=1.0.0=py37h99015e2_1
- multidict=4.7.5=py37h8f50634_1
- multipledispatch=0.6.0=py_0
- munch=2.5.0=py_0
- nbconvert=5.6.1=py37hc8dfbb8_1
- nbformat=5.0.7=py_0
- nccl=2.5.7.1=hc6a2c23_0
- ncurses=6.2=he6710b0_1
- networkx=2.4=py_1
- nodejs=13.13.0=hf5d1a2b_0
- notebook=6.0.3=py37hc8dfbb8_1
- numba=0.50.1=py37h0da4684_1
- numpy=1.19.0=py37h8960a57_0
- olefile=0.46=py_0
- openjpeg=2.3.1=h981e76c_3
- openssl=1.1.1g=h516909a_0
- packaging=20.4=pyh9f0ad1d_0
- pandas=1.0.5=py37h0da4684_0
- pandoc=2.10=h14c3975_0
- pandocfilters=1.4.2=py_1
- panel=0.9.7=py_0
- param=1.9.3=py_0
- parquet-cpp=1.5.1=2
- parso=0.7.0=pyh9f0ad1d_0
- partd=1.1.0=py_0
- pcre=8.44=he1b5a44_0
- pexpect=4.8.0=py37hc8dfbb8_1
- pickle5=0.0.11=py37h8f50634_0
- pickleshare=0.7.5=py37hc8dfbb8_1001
- pillow=7.2.0=py37h718be6c_1
- pip=20.1.1=py37_1
- pixman=0.38.0=h516909a_1003
- poppler=0.87.0=h4190859_1
- poppler-data=0.4.9=1
- postgresql=12.3=h8573dbc_0
- proj=7.0.0=h966b41f_5
- prometheus_client=0.8.0=pyh9f0ad1d_0
- prompt-toolkit=3.0.5=py_1
- psutil=5.7.2=py37h8f50634_0
- pthread-stubs=0.4=h14c3975_1001
- ptyprocess=0.6.0=py_1001
- py-xgboost=1.1.0dev.rapidsai0.15=cuda10.2py37_1
- pyarrow=0.17.1=py37h1234567_11_cuda
- pycparser=2.20=pyh9f0ad1d_2
- pyct=0.4.6=py_0
- pyct-core=0.4.6=py_0
- pydeck=0.3.1=pyh9f0ad1d_0
- pyee=7.0.2=pyh9f0ad1d_0
- pygments=2.6.1=py_0
- pynvml=8.0.4=py_1
- pyopenssl=19.1.0=py_1
- pyparsing=2.4.7=pyh9f0ad1d_0
- pyppeteer=0.0.25=py_1
- pyproj=2.6.1.post1=py37h34dd122_0
- pyrsistent=0.16.0=py37h8f50634_0
- pysocks=1.7.1=py37hc8dfbb8_1
- python=3.7.7=hcff3b4d_5
- python-dateutil=2.8.1=py_0
- python_abi=3.7=1_cp37m
- pytz=2020.1=pyh9f0ad1d_0
- pyviz_comms=0.7.6=pyh9f0ad1d_0
- pywavelets=1.1.1=py37h03ebfcd_1
- pyyaml=5.3.1=py37h8f50634_0
- pyzmq=19.0.1=py37hac76be4_0
- rapids=0.15.0a200721=cuda10.2_py37_gf0c1bc3_99
- rapids-xgboost=0.15.0a200721=cuda10.2_py37_gf0c1bc3_99
- re2=2020.07.06=he1b5a44_1
- readline=8.0=h7b6447c_0
- requests=2.24.0=pyh9f0ad1d_0
- rmm=0.15.0a200721=py37_ge2c0e95_368
- rtree=0.9.4=py37h8526d28_1
- scikit-image=0.17.2=py37h0da4684_1
- scikit-learn=0.23.1=py37h8a51577_0
- scipy=1.5.1=py37ha3d9a3c_0
- send2trash=1.5.0=py_0
- setuptools=49.2.0=py37_0
- shapely=1.7.0=py37hc88ce51_3
- simpervisor=0.3=py_1
- six=1.15.0=pyh9f0ad1d_0
- snappy=1.1.8=he1b5a44_3
- sortedcontainers=2.2.2=pyh9f0ad1d_0
- spdlog=1.7.0=hc9558a2_0
- sqlite=3.32.3=h62c20be_0
- tbb=2020.1=hc9558a2_0
- tblib=1.6.0=py_0
- terminado=0.8.3=py37hc8dfbb8_1
- testpath=0.4.4=py_0
- threadpoolctl=2.1.0=pyh5ca1d4c_0
- thrift-cpp=0.13.0=h62aa4f2_2
- tifffile=2020.7.17=py_0
- tiledb=1.7.7=h8efa9f0_3
- tk=8.6.10=hbc83047_0
- toolz=0.10.0=py_0
- tornado=6.0.4=py37h8f50634_1
- tqdm=4.48.0=pyh9f0ad1d_0
- traitlets=4.3.3=py37hc8dfbb8_1
- treelite=0.92=py37h023e13c_2
- typing_extensions=3.7.4.2=py_0
- tzcode=2020a=h516909a_0
- ucx=1.8.1+g6b29558=cuda10.2_0
- ucx-py=0.15.0a200721+g6b29558=py37_142
- urllib3=1.25.9=py_0
- wcwidth=0.2.5=pyh9f0ad1d_0
- webencodings=0.5.1=py_1
- websockets=8.1=py37h8f50634_1
- wheel=0.34.2=py37_0
- widgetsnbextension=3.5.1=py37hc8dfbb8_1
- xarray=0.16.0=py_0
- xerces-c=3.2.2=h8412b87_1004
- xgboost=1.1.0dev.rapidsai0.15=cuda10.2py37_1
- xorg-kbproto=1.0.7=h14c3975_1002
- xorg-libice=1.0.10=h516909a_0
- xorg-libsm=1.2.3=h84519dc_1000
- xorg-libx11=1.6.9=h516909a_0
- xorg-libxau=1.0.9=h14c3975_0
- xorg-libxdmcp=1.1.3=h516909a_0
- xorg-libxext=1.3.4=h516909a_0
- xorg-libxrender=0.9.10=h516909a_1002
- xorg-renderproto=0.11.1=h14c3975_1002
- xorg-xextproto=7.3.0=h14c3975_1002
- xorg-xproto=7.0.31=h14c3975_1007
- xz=5.2.5=h7b6447c_0
- yaml=0.2.5=h516909a_0
- yarl=1.3.0=py37h516909a_1000
- zeromq=4.3.2=he1b5a44_2
- zfp=0.5.5=he1b5a44_1
- zict=2.0.0=py_0
- zipp=3.1.0=py_0
- zlib=1.2.11=h7b6447c_3
- zstd=1.4.5=h6597ccf_1
- pip:
- alembic==1.4.2
- azure-core==1.7.0
- azure-storage-blob==12.3.2
- cliff==3.3.0
- cmaes==0.5.1
- cmd2==1.2.1
- colorama==0.4.3
- colorlog==4.1.0
- databricks-cli==0.11.0
- docker==4.2.2
- flask==1.1.2
- gitdb==4.0.5
- gitpython==3.1.7
- gorilla==0.3.0
- gunicorn==20.0.4
- isodate==0.6.0
- itsdangerous==1.1.0
- json5==0.9.5
- jupyterlab==2.2.0
- jupyterlab-server==1.2.0
- lightgbm==2.3.1
- mako==1.1.3
- ml-metrics==0.1.4
- mlflow==1.10.0
- msrest==0.6.17
- oauthlib==3.1.0
- optuna==1.5.0
- pbr==5.4.5
- prettytable==0.7.2
- prometheus-flask-exporter==0.15.0
- protobuf==3.12.2
- pyperclip==1.8.0
- python-editor==1.0.4
- querystring-parser==1.2.4
- requests-oauthlib==1.3.0
- smmap==3.0.4
- sqlalchemy==1.3.13
- sqlparse==0.3.1
- stevedore==3.2.0
- tabulate==0.8.7
- treelite-runtime==0.92
- websocket-client==0.57.0
- werkzeug==1.0.1
- xfeat==0.1.0
| 0 |
rapidsai_public_repos/cloud-ml-examples | rapidsai_public_repos/cloud-ml-examples/dask/README.md | # RAPIDS Hyperparameter Optimization (HPO) with Dask ML
[Dask-ML](https://ml.dask.org/) provides machine learning utilities
built on top of the scalable Dask platform. Dask already offers
[first-class integration with RAPIDS](https://rapids.ai/dask.html),
and Dask-ML is no exception.
The Dask-ML [hyperparameter search
tools](https://ml.dask.org/hyper-parameter-search.html) make it easy
to take advantage of grid search, randomized search, or hyperband HPO
algorithms. It particularly excels at incorporating cross-validation
into the HPO process for more stable accuracy estimates and at
allowing intelligent reuse of intermediate results from the Dask task
graph.
## RAPIDS + Dask-ML sample notebooks
This sample notebook shows how to use Ray Tune to optimize XGBoost and
cuML Random Forest classifiers over a large dataset of airline arrival
times. By design, it is very similar to the RAPIDS examples provided
for other cloud and bring-your-own-cloud HPO offerings.
You need both Jupyter and RAPIDS 0.13 or later installed to begin. See
https://rapids.ai/start.html for instructions. We recommend using the 21.06
packages for the latest updates. Dask-ML can be installed via conda or pip, following the instructions from: https://ml.dask.org/install.html.
## RAPIDS + Dask-Kubernetes sample notebooks
Dask_cuML_Exploration and Dask_cuML_Exploration_Full provide a template
for launching a dask cluster on top of your kubernetes environment, loading
the NYC-Taxi dataset, and generating performance metrics for the available
cuML Multi-Node Multi-GPU (MNMG) algorithms in your environment.
See: [Dask-Kubernetes](https://kubernetes.dask.org/en/latest/) and cuML's
[API documentation](https://docs.rapids.ai/api/cuml/stable/api.html#multi-node-multi-gpu-algorithms) for additional information.
| 0 |
rapidsai_public_repos/cloud-ml-examples/dask | rapidsai_public_repos/cloud-ml-examples/dask/notebooks/HPO_demo.ipynb | import warnings
warnings.filterwarnings('ignore') # Reduce number of messages/warnings displayedimport time
import numpy as np
import cupy as cp
import pandas as pd
import cudf
import cuml
import rmm
import xgboost as xgb
import sklearn.model_selection as sk
import dask_ml.model_selection as dcv
from dask.distributed import Client, wait
from dask_cuda import LocalCUDACluster
from sklearn import datasets
from sklearn.metrics import make_scorer
from sklearn.metrics import accuracy_score as sk_acc
from cuml.neighbors import KNeighborsClassifier
from cuml.ensemble import RandomForestClassifier
from cuml.model_selection import train_test_split
from cuml.metrics.accuracy import accuracy_score
import os
from urllib.request import urlretrieve
import gzipcluster = LocalCUDACluster(dashboard_address="127.0.0.1:8005")
client = Client(cluster)
clientdata_dir = '~/rapids_hpo/data/'
file_name = 'airlines.parquet'
parquet_name = os.path.join(data_dir, file_name)def prepare_dataset(use_full_dataset=False):
global file_path, data_dir
if use_full_dataset:
url = 'https://rapidsai-cloud-ml-sample-data.s3-us-west-2.amazonaws.com/airline_20000000.parquet'
else:
url = 'https://rapidsai-cloud-ml-sample-data.s3-us-west-2.amazonaws.com/airline_small.parquet'
if os.path.isfile(parquet_name):
print(f" > File already exists. Ready to load at {parquet_name}")
else:
# Ensure folder exists
os.makedirs(data_dir, exist_ok=True)
def data_progress_hook(block_number, read_size, total_filesize):
if (block_number % 1000) == 0:
print(
f" > percent complete: { 100 * ( block_number * read_size ) / total_filesize:.2f}\r",
end="",
)
return
urlretrieve(
url= url,
filename=parquet_name,
reporthook=data_progress_hook,
)
print(f" > Download complete {file_name}")
input_cols = ["Year", "Month", "DayofMonth", "DayofWeek", "CRSDepTime", "CRSArrTime",
"UniqueCarrier", "FlightNum", "ActualElapsedTime", "Origin", "Dest",
"Distance", "Diverted"]
dataset = cudf.read_parquet(parquet_name)
# encode categoricals as numeric
for col in dataset.select_dtypes(["object"]).columns:
dataset[col] = dataset[col].astype("category").cat.codes.astype(np.int32)
# cast all columns to int32
for col in dataset.columns:
dataset[col] = dataset[col].astype(np.float32) # needed for random forest
# put target/label column first [ classic XGBoost standard ]
output_cols = ["ArrDelayBinary"] + input_cols
dataset = dataset.reindex(columns=output_cols)
return datasetdf = prepare_dataset()import time
from contextlib import contextmanager
# Helping time blocks of code
@contextmanager
def timed(txt):
t0 = time.time()
yield
t1 = time.time()
print("%32s time: %8.5f" % (txt, t1 - t0))# Define some default values to make use of across the notebook for a fair comparison
N_FOLDS = 5
N_ITER = 25label = 'ArrDelayBinary'X_train, X_test, y_train, y_test = train_test_split(df, label,
test_size=0.2)X_cpu = X_train.to_pandas()
y_cpu = y_train.to_numpy()
X_test_cpu = X_test.to_pandas()
y_test_cpu = y_test.to_numpy()def accuracy_score_wrapper(y, y_hat):
"""
A wrapper function to convert labels to float32,
and pass it to accuracy_score.
Params:
- y: The y labels that need to be converted
- y_hat: The predictions made by the model
"""
y = y.astype("float32") # cuML RandomForest needs the y labels to be float32
return accuracy_score(y, y_hat, convert_dtype=True)
accuracy_wrapper_scorer = make_scorer(accuracy_score_wrapper)
cuml_accuracy_scorer = make_scorer(accuracy_score, convert_dtype=True)def do_HPO(model, gridsearch_params, scorer, X, y, mode='gpu-Grid', n_iter=10):
"""
Perform HPO based on the mode specified
mode: default gpu-Grid. The possible options are:
1. gpu-grid: Perform GPU based GridSearchCV
2. gpu-random: Perform GPU based RandomizedSearchCV
n_iter: specified with Random option for number of parameter settings sampled
Returns the best estimator and the results of the search
"""
if mode == 'gpu-grid':
print("gpu-grid selected")
clf = dcv.GridSearchCV(model,
gridsearch_params,
cv=N_FOLDS,
scoring=scorer)
elif mode == 'gpu-random':
print("gpu-random selected")
clf = dcv.RandomizedSearchCV(model,
gridsearch_params,
cv=N_FOLDS,
scoring=scorer,
n_iter=n_iter)
else:
print("Unknown Option, please choose one of [gpu-grid, gpu-random]")
return None, None
res = clf.fit(X, y)
print("Best clf and score {} {}\n---\n".format(res.best_estimator_, res.best_score_))
return res.best_estimator_, resdef print_acc(model, X_train, y_train, X_test, y_test, mode_str="Default"):
"""
Trains a model on the train data provided, and prints the accuracy of the trained model.
mode_str: User specifies what model it is to print the value
"""
y_pred = model.fit(X_train, y_train).predict(X_test)
score = accuracy_score(y_pred, y_test.astype('float32'), convert_dtype=True)
print("{} model accuracy: {}".format(mode_str, score))X_train.shapemodel_gpu_xgb_ = xgb.XGBClassifier(tree_method='gpu_hist')
print_acc(model_gpu_xgb_, X_train, y_cpu, X_test, y_test_cpu)# For xgb_model
model_gpu_xgb = xgb.XGBClassifier(tree_method='gpu_hist')
# More range
params_xgb = {
"max_depth": np.arange(start=3, stop = 12, step = 3), # Default = 6
"alpha" : np.logspace(-3, -1, 5), # default = 0
"learning_rate": [0.05, 0.1, 0.15], #default = 0.3
"min_child_weight" : np.arange(start=2, stop=10, step=3), # default = 1
"n_estimators": [100, 200, 1000]
}mode = "gpu-random"
with timed("XGB-"+mode):
res, results = do_HPO(model_gpu_xgb,
params_xgb,
cuml_accuracy_scorer,
X_train,
y_cpu,
mode=mode,
n_iter=N_ITER)
print("Searched over {} parameters".format(len(results.cv_results_['mean_test_score'])))print_acc(res, X_train, y_cpu, X_test, y_test_cpu, mode_str=mode)mode = "gpu-grid"
with timed("XGB-"+mode):
res, results = do_HPO(model_gpu_xgb,
params_xgb,
cuml_accuracy_scorer,
X_train,
y_cpu,
mode=mode)
print("Searched over {} parameters".format(len(results.cv_results_['mean_test_score'])))print_acc(res, X_train, y_cpu, X_test, y_test_cpu, mode_str=mode)from cuml.experimental.hyperopt_utils import plotting_utilsplotting_utils.plot_search_results(results)df_gridsearch = pd.DataFrame(results.cv_results_)
plotting_utils.plot_heatmap(df_gridsearch, "param_max_depth", "param_n_estimators")## Random Forest
model_rf_ = RandomForestClassifier()
params_rf = {
"max_depth": np.arange(start=3, stop = 15, step = 2), # Default = 6
"max_features": [0.1, 0.50, 0.75, 'auto'], #default = 0.3
"n_estimators": [100, 200, 500, 1000]
}
for col in X_train.columns:
X_train[col] = X_train[col].astype('float32')
y_train = y_train.astype("int32")print("Default acc: ",accuracy_score(model_rf_.fit(X_train, y_train).predict(X_test), y_test))mode = "gpu-random"
model_rf = RandomForestClassifier()
with timed("RF-"+mode):
res, results = do_HPO(model_rf,
params_rf,
cuml_accuracy_scorer,
X_train,
y_cpu,
mode=mode,
n_iter = N_ITER)
print("Searched over {} parameters".format(len(results.cv_results_['mean_test_score'])))print("Improved acc: ",accuracy_score(res.predict(X_test), y_test))df_gridsearch = pd.DataFrame(results.cv_results_)
plotting_utils.plot_heatmap(df_gridsearch, "param_max_depth", "param_n_estimators") | 0 |
rapidsai_public_repos/cloud-ml-examples/dask | rapidsai_public_repos/cloud-ml-examples/dask/kubernetes/Dask_cuML_Exploration.ipynb | import certifi
import cudf
import cuml
import cupy as cp
import gcsfs
import json
import numpy as np
import os
import pandas as pd
import random
import seaborn as sns
import time
import uuid
import yaml
from collections import OrderedDict
from functools import partial
from math import cos, sin, asin, sqrt, pi
from tqdm import tqdm
from typing import Optional
import dask
import dask.array as da
import dask.dataframe as dd
import dask_cudf
from dask_kubernetes import KubeCluster, make_pod_from_dict
from dask.distributed import Client, WorkerPlugin, wait, progress, get_worker
from cuml import ForestInference
def create_pod_from_yaml(yaml_file):
with open(yaml_file, 'r') as reader:
d = yaml.safe_load(reader)
d = dask.config.expand_environment_variables(d)
return make_pod_from_dict(d)
def build_worker_and_scheduler_pods(sched_spec, worker_spec):
assert os.path.isfile(sched_spec)
assert os.path.isfile(worker_spec)
sched_pod = create_pod_from_yaml(sched_spec)
worker_pod = create_pod_from_yaml(worker_spec)
return sched_pod, worker_pod
dask.config.set({"logging.kubernetes": "info",
"logging.distributed": "info",
"kubernetes.scheduler-service-type": "LoadBalancer",
"kubernetes.idle-timeout": None,
"kubernetes.scheduler-service-wait-timeout": 3600,
"kubernetes.deploy-mode": "remote",
"kubernetes.logging": "info",
"distributed.logging": "info",
"distributed.scheduler.idle-timeout": None,
"distributed.scheduler.locks.lease-timeout": None,
"distributed.comm.timeouts.connect": 3600,
"distributed.comm.tls.ca-file": certifi.where()})
sched_spec_path = "./specs/sched-spec.yaml"
worker_spec_path = "./specs/worker-spec.yaml"
sched_pod, worker_pod = build_worker_and_scheduler_pods(sched_spec=sched_spec_path,
worker_spec=worker_spec_path)cluster = KubeCluster(pod_template=worker_pod,
scheduler_pod_template=sched_pod)client = Client(cluster)
scheduler_address = cluster.scheduler_addressdef scale_workers(client, n_workers, timeout=300):
client.cluster.scale(n_workers)
m = len(client.has_what().keys())
start = end = time.perf_counter_ns()
while ((m != n_workers) and (((end - start) / 1e9) < timeout) ):
time.sleep(5)
m = len(client.has_what().keys())
end = time.perf_counter_ns()
if (((end - start) / 1e9) >= timeout):
raise RuntimeError(f"Failed to rescale cluster in {timeout} sec."
"Try increasing timeout for very large containers, and verify available compute resources.")scale_workers(client, n_workers=8 timeout=800)class SimpleTimer:
def __init__(self):
self.start = None
self.end = None
self.elapsed = None
def __enter__(self):
self.start = time.perf_counter_ns()
return self
def __exit__(self, exc_type, exc_val, exc_tb):
self.end = time.perf_counter_ns()
self.elapsed = self.end - self.start
def construct_worker_pool(client, n_workers, auto_scale=False, timeout=300):
workers = [w for w in client.has_what().keys()]
if (len(workers) < n_workers):
if (auto_scale):
scale_workers(client=client, n_workers=n_workers, timeout=timeout)
workers = [w for w in client.has_what().keys()]
else:
print("Attempt to construct worker pool larger than available worker set, and auto_scale is False."
" Returning entire pool.")
else:
workers = random.sample(population=workers, k=n_workers)
return workers
def estimate_df_rows(client, files, storage_opts={}, testpct=0.01):
workers = client.has_what().keys()
est_size = 0
for file in files:
if (file.endswith('.csv')):
df = dask_cudf.read_csv(file, npartitions=len(workers), storage_options=storage_opts)
elif (file.endswith('.parquet')):
df = dask_cudf.read_parquet(file, npartitions=len(workers))
# Select only the index column from our subsample
est_size += (df.sample(frac=testpct).iloc[:,0].shape[0] / testpct).compute()
del df
return est_sizedef clean(df_part, remap, must_haves):
"""
This function performs the various clean up tasks for the data
and returns the cleaned dataframe.
"""
tmp = {col:col.strip().lower() for col in list(df_part.columns)}
df_part = df_part.rename(columns=tmp)
# rename using the supplied mapping
df_part = df_part.rename(columns=remap)
# iterate through columns in this df partition
for col in df_part.columns:
# drop anything not in our expected list
if col not in must_haves:
df_part = df_part.drop(col, axis=1)
continue
# fixes datetime error found by Ty Mckercher and fixed by Paul Mahler
if df_part[col].dtype == 'object' and col in ['pickup_datetime', 'dropoff_datetime']:
df_part[col] = df_part[col].astype('datetime64[ms]')
continue
# if column was read as a string, recast as float
if df_part[col].dtype == 'object':
df_part[col] = df_part[col].astype('float32')
else:
# downcast from 64bit to 32bit types
# Tesla T4 are faster on 32bit ops
if 'int' in str(df_part[col].dtype):
df_part[col] = df_part[col].astype('int32')
if 'float' in str(df_part[col].dtype):
df_part[col] = df_part[col].astype('float32')
df_part[col] = df_part[col].fillna(-1)
return df_part
def coalesce_taxi_data(fraction, random_state):
base_path = 'gcs://anaconda-public-data/nyc-taxi/csv'
# list of column names that need to be re-mapped
remap = {}
remap['tpep_pickup_datetime'] = 'pickup_datetime'
remap['tpep_dropoff_datetime'] = 'dropoff_datetime'
remap['ratecodeid'] = 'rate_code'
#create a list of columns & dtypes the df must have
must_haves = {
'pickup_datetime': 'datetime64[ms]',
'dropoff_datetime': 'datetime64[ms]',
'passenger_count': 'int32',
'trip_distance': 'float32',
'pickup_longitude': 'float32',
'pickup_latitude': 'float32',
'rate_code': 'int32',
'dropoff_longitude': 'float32',
'dropoff_latitude': 'float32',
'fare_amount': 'float32'
}
# apply a list of filter conditions to throw out records with missing or outlier values
query_fragments = [
'fare_amount > 0 and fare_amount < 500',
'passenger_count > 0 and passenger_count < 6',
'pickup_longitude > -75 and pickup_longitude < -73',
'dropoff_longitude > -75 and dropoff_longitude < -73',
'pickup_latitude > 40 and pickup_latitude < 42',
'dropoff_latitude > 40 and dropoff_latitude < 42'
]
valid_months_2016 = [str(x).rjust(2, '0') for x in range(1, 7)]
valid_files_2016 = [f'{base_path}/2016/yellow_tripdata_2016-{month}.csv' for month in valid_months_2016]
df_2014_fractional = dask_cudf.read_csv(f'{base_path}/2014/yellow_*.csv', chunksize=25e6).sample(
frac=fraction, random_state=random_state)
df_2014_fractional = clean(df_2014_fractional, remap, must_haves)
df_2015_fractional = dask_cudf.read_csv(f'{base_path}/2015/yellow_*.csv', chunksize=25e6).sample(
frac=fraction, random_state=random_state)
df_2015_fractional = clean(df_2015_fractional, remap, must_haves)
df_2016_fractional = dask_cudf.read_csv(valid_files_2016, chunksize=25e6).sample(
frac=fraction, random_state=random_state)
df_2016_fractional = clean(df_2016_fractional, remap, must_haves)
df_taxi = dask.dataframe.multi.concat([df_2014_fractional, df_2015_fractional, df_2016_fractional])
df_taxi = df_taxi.query(' and '.join(query_fragments))
return df_taxidef persist_train_infer_split(client, df, response_dtype, response_id, infer_frac=1.0):
workers = client.has_what().keys()
infer_frac = max(0, min(infer_frac, 1.0))
df_train = df
X_train = df_train[df.columns.difference([response_id])].astype(np.float32)
y_train = df_train[response_id].astype(response_dtype)
with dask.annotate(workers=set(workers)):
X_train, y_train = client.persist(
collections=[X_train, y_train])
if (infer_frac != 1.0):
df_split = df.sample(frac=infer_frac)
X_infer = df_split[df.columns.difference([response_id])].astype(np.float32)
y_infer = df_split[response_id].astype(response_dtype)
with dask.annotate(workers=set(workers)):
X_infer, y_infer = client.persist(
collections=[X_infer, y_infer])
wait([X_train, y_train, X_infer, y_infer])
else:
X_infer = X_train
y_infer = y_train
wait([X_train, y_train])
return X_train, y_train, X_train, y_train
def mortgage_parquet_loader(client, response_dtype=np.float32, fraction=1.0, infer_frac=1.0, random_state=0):
response_id = 'foreclosure_costs'
km_fields = [
'loan_id', 'interest_rate', 'current_actual_upb', 'loan_age', 'remaining_months_to_legal_maturity',
'msa', 'current_loan_delinquency_status',
'prop_preservation_and_repair_costs', 'asset_recovery_costs', 'misc_holding_expenses', 'holding_taxes',
'net_sale_proceeds', 'credit_enhancement_proceeds', 'repurchase_make_whole_proceeds',
'other_foreclosure_proceeds', 'non_interest_bearing_upb', 'principal_forgiveness_upb',
'foreclosure_principal_write_off_amount', 'foreclosure_costs'
]
mort_df = dask_cudf.read_parquet(YOUR_MORTGAGE_DATA_PATH, storage_options=YOUR_STORAGE_OPTS).sample(
frac=fraction, random_state=random_state).fillna(value=0.0)
mort_df = mort_df[km_fields]
return persist_train_infer_split(client, mort_df, response_dtype, response_id, infer_frac)
def taxi_csv_data_loader(client, response_dtype=np.float32, fraction=1.0, random_state=0):
response_id = 'fare_amount'
workers = client.has_what().keys()
km_fields = ['passenger_count', 'trip_distance', 'pickup_longitude',
'pickup_latitude', 'rate_code', 'dropoff_longitude',
'dropoff_latitude', 'fare_amount']
taxi_df = coalesce_taxi_data(fraction=fraction, random_state=random_state)
taxi_df = taxi_df[km_fields]
X = taxi_df[taxi_df.columns.difference([response_id])].astype(np.float32)
y = taxi_df[response_id].astype(response_dtype)
with dask.annotate(workers=set(workers)):
taxi_df = client.persist(collections=taxi_df)
with dask.annotate(workers=set(workers)):
X = client.persist(collections=X)
with dask.annotate(workers=set(workers)):
y = client.persist(collections=y)
wait([taxi_df, X, y])
return taxi_df, X, y
def taxi_parquet_data_loader(client, response_dtype=np.float32, fraction=1.0, infer_frac=1.0, random_state=0):
# list of column names that need to be re-mapped
remap = {}
remap['tpep_pickup_datetime'] = 'pickup_datetime'
remap['tpep_dropoff_datetime'] = 'dropoff_datetime'
remap['ratecodeid'] = 'rate_code'
#create a list of columns & dtypes the df must have
must_haves = {
'pickup_datetime': 'datetime64[ms]',
'dropoff_datetime': 'datetime64[ms]',
'passenger_count': 'int32',
'trip_distance': 'float32',
'pickup_longitude': 'float32',
'pickup_latitude': 'float32',
'rate_code': 'int32',
'dropoff_longitude': 'float32',
'dropoff_latitude': 'float32',
'fare_amount': 'float32'
}
# apply a list of filter conditions to throw out records with missing or outlier values
query_fragments = [
'fare_amount > 0 and fare_amount < 500',
'passenger_count > 0 and passenger_count < 6',
'pickup_longitude > -75 and pickup_longitude < -73',
'dropoff_longitude > -75 and dropoff_longitude < -73',
'pickup_latitude > 40 and pickup_latitude < 42',
'dropoff_latitude > 40 and dropoff_latitude < 42'
]
workers = client.has_what().keys()
taxi_parquet_path = "gs://anaconda-public-data/nyc-taxi/nyc.parquet"
response_id = 'fare_amount'
fields = ['passenger_count', 'trip_distance', 'pickup_longitude', 'pickup_latitude', 'rate_code',
'dropoff_longitude', 'dropoff_latitude', 'fare_amount']
taxi_df = dask_cudf.read_parquet(taxi_parquet_path, npartitions=len(workers))
taxi_df = clean(taxi_df, remap, must_haves)
taxi_df = taxi_df.query(' and '.join(query_fragments))
taxi_df = taxi_df[fields]
return persist_train_infer_split(client, taxi_df, response_dtype, response_id, infer_frac)def record_elapsed_timings_to_df(df, timings, record_template, type, columns, write_to=None):
records = [dict(record_template, **{"sample_index": i,
"elapsed": elapsed,
"type": type})
for i, elapsed in enumerate(timings)]
df = df.append(other=records, ignore_index=True)
if (write_to):
df.to_csv(write_to, columns=columns)
return df
def collect_load_time_samples(load_func, count, return_final_sample=True, verbose=False):
timings = []
for m in tqdm(range(count)):
with SimpleTimer() as timer:
X_train, y_train, X_infer, y_infer = load_func()
timings.append(timer.elapsed)
if (return_final_sample):
return X_train, y_train, X_infer, y_infer, timings
return None, None, None, timings
def collect_func_time_samples(func, count, verbose=False):
timings = []
for k in tqdm(range(count)):
with SimpleTimer() as timer:
func()
timings.append(timer.elapsed)
return timings
def sweep_fit_func(model, func_id, require_compute, X, y, xy_fit, count):
_fit_func_attr = getattr(model, func_id)
if (require_compute):
if (xy_fit):
fit_func = partial(lambda X, y: _fit_func_attr(X, y).compute(), X, y)
else:
fit_func = partial(lambda X: _fit_func_attr(X).compute(), X)
else:
if (xy_fit):
fit_func = partial(_fit_func_attr, X, y)
else:
fit_func = partial(_fit_func_attr, X)
return collect_func_time_samples(func=fit_func, count=count)
def sweep_predict_func(model, func_id, require_compute, X, count):
_predict_func_attr = getattr(model, func_id)
predict_func = partial(lambda X: _predict_func_attr(X).compute(), X)
return collect_func_time_samples(func=predict_func, count=count)
def performance_sweep(client, model, data_loader, hardware_type, worker_counts=[1], samples=1, load_samples=1, max_data_frac=1.0,
predict_frac=1.0, scaling_type='weak', xy_fit=True, fit_requires_compute=False, update_workers_in_kwargs=True,
response_dtype=np.float32, out_path='./perf_sweep.csv', append_to_existing=False, model_name=None, infer_with_fil=False,
fit_func_id="fit", predict_func_id="predict", scaling_denom=None, post_fit_handler=None, model_args={}, model_kwargs={}):
"""
Primary performance sweep entrypoint.
Parameters
------------
client: DASK client associated with the cluster we're interesting in collecting performance data for.
model: Model object on which to gather performance data. This will be created and destroyed,
once for each element of 'worker_counts'
data_loader: arbitrary data loading function that will be called to load the appropriate testing data.
Function that is responsible for loading and returning the data to be used for a given performance run. Function
signature must accept (client, fraction, and random_state). Client should be used to distribute data, and loaders
should utilize fraction and random_state with dask's dataframe.sample method to allow for control of how much data
is loaded.
When called, its return value should be of the form: df, X, y, where df is the full dask_cudf dataframe, X is a
dask_cudf dataframe which contains all explanatory variables that will be passed to the 'fit' function, and y is a
dask_cudf series or dataframe that contains response variables which should be passed to fit/predict as fit(X, y)
hardware_type: indicates the core hardware the current sweep is running on. ex. 'T4', 'V100', 'A100'
worker_counts: List indicating the number of workers that should be swept. Ex [1, 2, 4]
worker counts must fit within the cluster associated with 'client', if the current DASK worker count is different
from what is requested on a given sweep, attempt to automatically scale the worker count. NOTE: this does not
mean we will scale the available cluster nodes, just the number of deployed worker pods.
samples: number of fit/predict samples to record per worker count
load_samples: number of times to sample data loads. This effectively times how long 'data_loader' runs.
max_data_frac: maximum fraction of data to return.
Strong scaling: each run will utilize max_data_frac data.
Weak scaling: each run will utilize (current worker count) / (max worker count) * max_data_frac data.
predict_frac: fraction of training data used to test inference
scaling_type: values can be 'weak' or 'strong' indicating the type of scaling sweep to perform.
xy_fit: indicates whether or not the model's 'fit' function is of the form (X, y), when xy_fit is False, we assume that
fit is of the form (X), as is the case with various unsupervised methods ex. KNN.
fit_requires_compute: False generally, set this to True if the model's 'fit' function requires a corresponding '.compute()'
call to execute the required work.
update_workers_in_kwargs: Some algorithms accept a 'workers' list, much like DASK, and will require their kwargs to have
workers populated. Setting this flag handles this automatically.
response_dtype: defaults to np.float32, some algorithms require another dtype, such as int32
out_path: path where performance data csv should be saved
append_to_existing: When true, append results to an existing csv, otherwise overwrite.
model_name: Override what we output as the model name
fit_func_id: Defaults to 'fit', only set this if the model has a non-standard naming.
predict_func_id: Defaults to 'predict', only set this if the model has a on-standard predict naming.
scaling_denom: (weak scaling) defaults to max(workers) if unset. Specifies the maximum worker count that weak scaling
should scale against. For example, when using 1 worker in a weak scaling sweep, the worker will attempt to
process a fraction of the total data equal to 1/scaling_denom
model_args: args that will be passed to the model's constructor
model_kwargs: keyword args that will be passed to the model's constructor
Returns
--------
"""
cols = ['n_workers', 'sample_index', 'elapsed', 'type', 'algorithm', 'scaling_type', 'data_fraction', 'hardware', 'trial_id']
perf_df = cudf.DataFrame(columns=cols)
if (append_to_existing):
try:
perf_df = cudf.read_csv(out_path)
except:
pass
model_name = model_name if model_name else str(model)
scaling_denom = scaling_denom if (scaling_denom is not None) else max(worker_counts)
max_data_frac = min(1.0, max_data_frac)
start_msg = f"Starting {scaling_type}-scaling performance sweep for:\n"
start_msg += f" model : {model_name}\n"
start_msg += f" data loader: {data_loader}.\n"
start_msg += f"Configuration\n"
start_msg += "==========================\n"
start_msg += f"{'Worker counts':<25} : {worker_counts}\n"
start_msg += f"{'Fit/Predict samples':<25} : {samples}\n"
start_msg += f"{'Data load samples':<25} : {load_samples}\n"
start_msg += f"- {'Max data fraction':<23} : {max_data_frac:0.2f}\n"
start_msg += f" - {'Train':<22} : {max_data_frac:0.2f}\n"
start_msg += f" - {'Infer':<22} : {predict_frac*max_data_frac:0.2f}\n"
start_msg += f"{'Model fit':<25} : {'X ~ y' if xy_fit else 'X'}\n"
start_msg += f"- {'Response DType':<23} : {response_dtype}\n"
start_msg += f"{'Writing results to':<25} : {out_path}\n"
start_msg += f"- {'Method':<23} : {'overwrite' if not append_to_existing else 'append'}\n"
print(start_msg, flush=True)
for n in worker_counts:
fraction = (n / scaling_denom) * max_data_frac if scaling_type == 'weak' else max_data_frac
fraction = min(1.0, fraction)
record_template = {
"n_workers": n,
"type": "predict",
"algorithm": model_name,
"scaling_type": scaling_type,
"data_fraction": fraction,
"hardware": hardware_type,
"trial_id": str(uuid.uuid4()),
}
scale_workers(client, n)
print(f"Sampling <{load_samples}> load times with {n} workers.",
f" With {fraction*100:0.1f} percent of total data", flush=True)
# Todo: add fractional selection passthrough
load_func = partial(data_loader, client=client, response_dtype=response_dtype, fraction=fraction, random_state=0)
X_train, y_train, X_infer, y_infer, load_timings = collect_load_time_samples(load_func=load_func, count=load_samples)
perf_df = record_elapsed_timings_to_df(df=perf_df, timings=load_timings, type='load',
record_template=record_template, columns=cols,write_to=out_path)
print(f"Finished loading <{load_samples}>, samples, to <{n}>",
f"workers with a mean time of {np.mean(load_timings)/1e9:0.4f} sec.", flush=True)
print(f"Sweeping {model_name} '{fit_func_id}' with <{n}> workers. Sampling",
f" <{samples}> times with {fraction*100:0.1f} percent of total data.", flush=True)
if (update_workers_in_kwargs and 'workers' in model_kwargs):
model_kwargs['workers'] = workers = list(client.has_what().keys())
m = model(*model_args, **model_kwargs)
if (fit_func_id):
fit_timings = sweep_fit_func(model=m, func_id=fit_func_id,
require_compute=fit_requires_compute,
X=X_train, y=y_train, xy_fit=xy_fit, count=samples)
perf_df = record_elapsed_timings_to_df(df=perf_df, timings=fit_timings, type='fit',
record_template=record_template, columns=cols, write_to=out_path)
print(f"Finished gathering <{samples}>, 'fit' samples using <{n}>",
f" workers, with a mean time of {np.mean(fit_timings)/1e9:0.4f} sec.", flush=True)
else:
print(f"Skipping fit sweep, fit_func_id is None")
if (post_fit_handler):
post_fit_handler(m)
if (predict_func_id):
print(f"Sweeping {model_name} '{predict_func_id}' with <{n}> workers."
f" Sampling <{samples}> times with {fraction*predict_frac*100:0.1f} percent of total data.", flush=True)
predict_timings = sweep_predict_func(model=m, func_id=predict_func_id,
require_compute=True, X=X_infer, count=samples)
perf_df = record_elapsed_timings_to_df(df=perf_df, timings=predict_timings, type='predict',
record_template=record_template, columns=cols, write_to=out_path)
print(f"Finished gathering <{samples}>, 'predict' samples using <{n}>",
f" workers, with a mean time of {np.mean(predict_timings)/1e9:0.4f} sec.",
flush=True)
else:
print(f"Skipping inference sweep. predict_func_id is None")def simple_ci(df, fields, groupby):
gbdf = df[fields].groupby(groupby).agg(['mean', 'std', 'count'])
ci = (1.96 + gbdf['elapsed']['std'] / np.sqrt(gbdf['elapsed']['count']))
ci_df = ci.reset_index()
ci_df['ci.low'] = gbdf['elapsed'].reset_index()['mean'] - ci_df[0]
ci_df['ci.high'] = gbdf['elapsed'].reset_index()['mean'] + ci_df[0]
return ci_df
def visualize_csv_data(csv_path, filter_query=None):
import pandas as pd
df = cudf.read_csv(csv_path)
fields = ['elapsed', 'elapsed_sec', 'type', 'n_workers', 'hardware', 'scaling_type']
groupby = ['n_workers', 'type', 'hardware', 'scaling_type']
df['elapsed_sec'] = df['elapsed']/1e9
ci_df = simple_ci(df, fields, groupby=groupby)
# Rescale to seconds
ci_df[['ci.low', 'ci.high']] = ci_df[['ci.low', 'ci.high']]/1e9
# Print confidence intervals
print(ci_df[['hardware', 'n_workers', 'type', 'ci.low', 'ci.high']][ci_df['type'] != 'load'])
sns.set_theme(style="whitegrid")
sns.set(rc={'figure.figsize':(20, 10)}, font_scale=2)
# Boxplots for elapsed time at each worker count.
# [df[fields].type != 'load']
plot_df = df[fields].to_pandas()
plot_df = plot_df.query("type != 'load'")
if (filter_query):
plot_df = plot_df.query(filter_query)
ax = sns.catplot(data=plot_df, x="n_workers", y="elapsed_sec",
col="type", row="scaling_type", hue="hardware", kind="box",
height=8, order=None)# Uncomment to test with Taxi Dataset
append_to_existing = True
samples = 5
load_samples = 1
worker_counts = [8]
scaling_denom = 8
hardware_type = None
max_data_frac = 0.75
scale_type = 'weak' # weak | strong
out_prefix = 'taxi_medium'
data_loader = taxi_parquet_data_loader
if (not hardware_type):
raise RuntimeError("Please specify the hardware type for this run! ex. (T4, V100, A100)")
sweep_kwargs = {
'append_to_existing': append_to_existing,
'samples': samples,
'load_samples': load_samples,
'worker_counts': worker_counts,
'scaling_denom': scaling_denom,
'hardware_type': hardware_type,
'data_loader': data_loader,
'max_data_frac': max_data_frac,
'scaling_type': scale_type
}taxi_parquet_path = ["gs://anaconda-public-data/nyc-taxi/nyc.parquet"]
estimated_rows = estimate_df_rows(client, files=taxi_parquet_path, testpct=0.0001)
print(estimated_rows)# Uncomment to sweep with the large Taxi Dataset
append_to_existing = True
samples = 5
load_samples = 1
worker_counts = [8,4,2]
scaling_denom = 8
hardware_type = None
max_data_frac = 0.75
scale_type = 'weak'
out_prefix = 'taxi_large'
data_loader = taxi_csv_data_loader
if (not hardware_type):
raise RuntimeError("Please specify the hardware type for this run! ex. (T4, V100, A100)")
sweep_kwargs = {
'append_to_existing': append_to_existing,
'samples': samples,
'load_samples': load_samples,
'worker_counts': worker_counts,
'scaling_denom': scaling_denom,
'hardware_type': hardware_type,
'data_loader': data_loader,
'max_data_frac': max_data_frac,
'scaling_type': scale_type
}append_to_existing = True
samples = 5
load_samples = 1
worker_counts = [4]
scaling_denom = 8
hardware_type = None
max_data_frac = 0.15
scale_type = 'weak'
out_prefix = 'mortgage_large'
data_loader = mortgage_parquet_loader
if (not hardware_type):
raise RuntimeError("Please specify the hardware type for this run! ex. (T4, V100, A100)")
sweep_kwargs = {
'append_to_existing': append_to_existing,
'samples': samples,
'load_samples': load_samples,
'worker_counts': worker_counts,
'scaling_denom': scaling_denom,
'hardware_type': hardware_type,
'data_loader': data_loader,
'max_data_frac': max_data_frac,
'scaling_type': scale_type
}remap = {}
remap['tpep_pickup_datetime'] = 'pickup_datetime'
remap['tpep_dropoff_datetime'] = 'dropoff_datetime'
remap['ratecodeid'] = 'rate_code'
#create a list of columns & dtypes the df must have
must_haves = {
'pickup_datetime': 'datetime64[ms]',
'dropoff_datetime': 'datetime64[ms]',
'passenger_count': 'int32',
'trip_distance': 'float32',
'pickup_longitude': 'float32',
'pickup_latitude': 'float32',
'rate_code': 'int32',
'dropoff_longitude': 'float32',
'dropoff_latitude': 'float32',
'fare_amount': 'float32'
}
# apply a list of filter conditions to throw out records with missing or outlier values
query_fragments = [
'fare_amount > 0 and fare_amount < 500',
'passenger_count > 0 and passenger_count < 6',
'pickup_longitude > -75 and pickup_longitude < -73',
'dropoff_longitude > -75 and dropoff_longitude < -73',
'pickup_latitude > 40 and pickup_latitude < 42',
'dropoff_latitude > 40 and dropoff_latitude < 42'
]
workers = client.has_what().keys()base_path = 'gcs://anaconda-public-data/nyc-taxi/csv'
with SimpleTimer() as timer_csv:
df_csv_2014 = dask_cudf.read_csv(f'{base_path}/2014/yellow_*.csv', chunksize=25e6, dtype={' tolls_amount': 'float64'})
df_csv_2014 = clean(df_csv_2014, remap, must_haves)
df_csv_2014 = df_csv_2014.query(' and '.join(query_fragments))
with dask.annotate(workers=set(workers)):
df_csv_2014 = client.persist(collections=df_csv_2014)
wait(df_csv_2014)
print(df_csv_2014.columns)
rows_csv = df_csv_2014.iloc[:,0].shape[0].compute()
print(f"CSV load took {timer_csv.elapsed/1e9} sec. For {rows_csv} rows of data => {rows_csv/(timer_csv.elapsed/1e9)} rows/sec")client.cancel(df_csv_2014)with SimpleTimer() as timer_parquet:
df_parquet = dask_cudf.read_parquet(f'gs://anaconda-public-data/nyc-taxi/nyc.parquet', chunksize=25e6)
df_parquet = clean(df_parquet, remap, must_haves)
df_parquet = df_parquet.query(' and '.join(query_fragments))
with dask.annotate(workers=set(workers)):
df_parquet = client.persist(collections=df_parquet)
wait(df_parquet)
print(df_parquet.columns)
rows_parquet = df_parquet.iloc[:,0].shape[0].compute()
print(f"Parquet load took {timer_parquet.elapsed/1e9} sec. For {rows_parquet} rows of data => {rows_parquet/(timer_parquet.elapsed/1e9)} rows/sec")client.cancel(df_parquet)speedup = (rows_parquet/(timer_parquet.elapsed/1e9))/(rows_csv/(timer_csv.elapsed/1e9))
print(speedup)from cuml.dask.ensemble import RandomForestRegressor
rf_kwargs = {
"workers": client.has_what().keys(),
"n_estimators": 200,
"max_depth": 6 # match xgboost's default
}
rf_csv_path = f"./{out_prefix}_random_forest_regression.csv"
performance_sweep(client=client, model=RandomForestRegressor,
**sweep_kwargs,
out_path=rf_csv_path,
response_dtype=np.int32,
model_kwargs=rf_kwargs)rf_csv_path = f"./{out_prefix}_random_forest_regression.csv"
visualize_csv_data(rf_csv_path)import xgboost as xgb
xg_args = [client]
xg_kwargs = {
'params': {
'tree_method': 'gpu_hist',
},
'num_boost_round': 100
}
xgb_csv_path = f'./{out_prefix}_xgb.csv'
def xgb_post_fit(model, gcs_model_path, project, token):
model.trained_model.save_model("./xgb_perf_sweep_model.model")
gcs_fs = gcsfs.core.GCSFileSystem(project=project, token=token)
# Push the model to central storage
gcs_fs.put("./xgb_perf_sweep_model.model", gcs_model_path)
client.run(attach_fil_to_worker, gcs_model_path, project, token,
wait=True)
def attach_fil_to_worker(gcs_model_path, project, token):
worker = get_worker()
gcs_fs = gcsfs.core.GCSFileSystem(project=project, token=token)
with gcs_fs.open(gcs_model_path, 'rb') as fmod:
data = fmod.read()
with open("./xgb_perf_sweep_model.model", 'wb') as wmod:
wmod.write(data)
worker.data["fil_model"] = ForestInference.load("./xgb_perf_sweep_model.model",
algo="BATCH_TREE_REORG",
output_class=False,
model_type='xgboost')
def predict_on_worker(X):
worker = get_worker()
worker.data['fil_model'].predict(X)
return
class XGBProxy():
"""
Create a simple API wrapper around XGBoost so that it supports our fit/predict workflow.
Parameters
-------------
data_loader: data loader object intended to be used by performance sweep.
"""
def __init__(self, data_loader):
self.args = []
self.kwargs = {}
self.data_loader = data_loader
self.trained_model = None
self.saved = False
self.fm = None
self.X_train = None
self.y_train = None
self.X_infer = None
self.y_infer = None
def loader(self, client, response_dtype, fraction, random_state):
"""
Wrap the data loader method so that it creates a DMatrix from the returned data.
"""
for df in [self.X_train, self.X_infer, self.y_train, self.y_infer]:
if df is not None:
del df
X_train, y_train, X_infer, y_infer = self.data_loader(client, response_dtype, fraction, random_state)
self.X_train = X_train
self.y_train = y_train
self.X_infer = X_infer
self.y_infer = y_infer
self.dmatrix = xgb.dask.DaskDMatrix(client, X_train, y_train)
return X_train, y_train, X_infer, y_infer
def __call__(self, *args, **kwargs):
"""
Acts as a pseudo init function which initializes our model args.
"""
self.args = args
self.kwargs = kwargs
return self
def fit(self, X):
"""
Wrap dask.train, and store the model on our proxy object.
"""
if (self.trained_model):
del self.trained_model
self.trained_model = xgb.dask.train(*self.args,
dtrain=self.dmatrix,
evals=[(self.dmatrix, 'train')],
**self.kwargs)['booster']
return self
def predict(self, X):
part_map = self.X_infer.map_partitions(predict_on_worker,
meta=pd.Series(dtype='float32'))
return part_map
_xgb_post_fit = partial(xgb_post_fit,
gcs_model_path=YOUR_GCS_MODEL_PATH,
project=YOUR_GCP_PROJECT,
token=YOUR_GCP_TOKEN)
xgb_proxy = XGBProxy(data_loader)
performance_sweep(client=client, model=xgb_proxy, data_loader=xgb_proxy.loader, hardware_type=hardware_type,
worker_counts=worker_counts,
samples=samples,
load_samples=load_samples,
max_data_frac=max_data_frac,
scaling_type=scale_type,
out_path=xgb_csv_path,
append_to_existing=append_to_existing,
update_workers_in_kwargs=False,
xy_fit=False,
scaling_denom = scaling_denom,
post_fit_handler=_xgb_post_fit,
model_args=xg_args,
model_kwargs=xg_kwargs)xgb_csv_path = f'./{out_prefix}_xgb.csv'
visualize_csv_data(xgb_csv_path) | 0 |
rapidsai_public_repos/cloud-ml-examples/dask | rapidsai_public_repos/cloud-ml-examples/dask/kubernetes/README.md | # Exploring cuML algorithms with dask-kubernetes
This guide aims to showcase a joint working example of [`cuML`](https://docs.rapids.ai/api/cuml/stable/),
[`dask-kubernetes`](https://kubernetes.dask.org/en/latest/index.html) and [Kubernetes cluster](https://kubernetes.io/).
## Prerequisite
A Kubernetes cluster capable of supporting 1 dask-scheduler and 8 dask-cuda-workers.
See `spec/sched-spec.yaml` and `spec/worker-spec.yaml` for resource requirements.
Visit [`rapidsai/deployment`](docs.rapids.ai/deployment) for setting up your cluster from various
cloud service provider (CSP).
## Launch Client
`Dockerfile` contains the image capable of running the notebooks in this folder.
Build the image:
```bash
docker build -t rapids-dask-kubernetes-client:22.06 .
```
Launch the container.
The container will automatically start a jupyter server and a dask dashboard.
Expose these services to local ports.
```bash
docker run --gpus all --rm -it --shm-size=1g --ulimit memlock=-1 \
-p 8888:8888 \
-p 8787:8787 \
-p 8786:8786 \
rapids-dask-kubernetes-client:22.06
```
> **note**
> The three ports exposed here are for jupyter-lab, dask-dashboard, dask-scheduler
> respectively.
## Execute Notebooks
Enter `cloud-ml-examples/dask/kubernetes` and explore the notebooks.
- `Dask_cuML_Exploration.ipynb`,
performs performance sweep of `RandomForesetRegressor` and XGBoost over fil model.
- `Dask_cuML_Exploration_Full.ipynb`,
extended version of above and performs performance sweep of more cuML APIs.
### Setup Cluster via dask-kubernetes
The first few cells of the notebooks launches a dask-cluster on your kubernetes cluster.
The scheduler and worker pods specifications are defined in `spec/sched-spec.yaml` and `spec/worker-spec.yaml`.
The specs are loaded via `create_pod_from_yaml` and passed to `KubeCluster`,
which will create the scheduler pod and services exposing the scheduler pods.
Worker pods are then created by `scale_worker` function.
| 0 |
rapidsai_public_repos/cloud-ml-examples/dask | rapidsai_public_repos/cloud-ml-examples/dask/kubernetes/Dask_cuML_Exploration_Full.ipynb | import certifi
import cudf
import cuml
import cupy as cp
import numpy as np
import os
import pandas as pd
import random
import seaborn as sns
import time
import yaml
from functools import partial
from math import cos, sin, asin, sqrt, pi
from tqdm import tqdm
from typing import Optional
import dask
import dask.array as da
import dask_cudf
from dask_kubernetes import KubeCluster, make_pod_from_dict
from dask.distributed import Client, WorkerPlugin, wait, progress
class SimpleTimer:
def __init__(self):
self.start = None
self.end = None
self.elapsed = None
def __enter__(self):
self.start = time.perf_counter_ns()
return self
def __exit__(self, exc_type, exc_val, exc_tb):
self.end = time.perf_counter_ns()
self.elapsed = self.end - self.start
def create_pod_from_yaml(yaml_file):
with open(yaml_file, 'r') as reader:
d = yaml.safe_load(reader)
d = dask.config.expand_environment_variables(d)
return make_pod_from_dict(d)
def build_worker_and_scheduler_pods(sched_spec, worker_spec):
assert os.path.isfile(sched_spec)
assert os.path.isfile(worker_spec)
sched_pod = create_pod_from_yaml(sched_spec)
worker_pod = create_pod_from_yaml(worker_spec)
return sched_pod, worker_pod
dask.config.set({"kubernetes.scheduler-service-type": "LoadBalancer",
"kubernetes.idle-timeout": None,
"kubernetes.scheduler-service-wait-timeout": 3600,
"kubernetes.deploy-mode": "remote",
"distributed.scheduler.idle-timeout": None,
"distributed.scheduler.locks.lease-timeout": None,
"distributed.comm.timeouts.connect": 3600,
"distributed.comm.tls.ca-file": certifi.where()})
sched_pod, worker_pod = build_worker_and_scheduler_pods(sched_spec='./specs/sched-spec.yaml',
worker_spec='./specs/worker-spec.yaml')cluster = KubeCluster(pod_template=worker_pod,
scheduler_pod_template=sched_pod)
client = Client(cluster)
scheduler_address = cluster.scheduler_addressn_workers = 4
cluster.scale(n_workers)def scale_workers(client, n_workers, timeout=300):
client.cluster.scale(n_workers)
m = len(client.has_what().keys())
start = end = time.perf_counter_ns()
while ((m != n_workers) and (((end - start) / 1e9) < timeout) ):
time.sleep(5)
m = len(client.has_what().keys())
end = time.perf_counter_ns()
if (((end - start) / 1e9) >= timeout):
raise RuntimeError(f"Failed to rescale cluster in {timeout} sec."
"Try increasing timeout for very large containers, and verify available compute resources.")
def construct_worker_pool(client, n_workers, auto_scale=False, timeout=300):
workers = [w for w in client.has_what().keys()]
if (len(workers) < n_workers):
if (auto_scale):
scale_workers(client=client, n_workers=n_workers, timeout=timeout)
workers = [w for w in client.has_what().keys()]
else:
print("Attempt to construct worker pool larger than available worker set, and auto_scale is False."
" Returning entire pool.")
else:
workers = random.sample(population=workers, k=n_workers)
return workers
def estimate_df_rows(client, files, storage_opts={}, testpct=0.01):
workers = client.has_what().keys()
est_size = 0
for file in files:
if (file.endswith('.csv')):
df = dask_cudf.read_csv(file, npartitions=len(workers), storage_options=storage_opts)
elif (file.endswith('.parquet')):
df = dask_cudf.read_parquet(file, npartitions=len(workers), storage_options=storage_opts)
# Select only the index column from our subsample
est_size += (df.sample(frac=testpct).iloc[:,0].shape[0] / testpct).compute()
del df
return est_sizedef clean(df_part, remap, must_haves):
"""
This function performs the various clean up tasks for the data
and returns the cleaned dataframe.
"""
tmp = {col:col.strip().lower() for col in list(df_part.columns)}
df_part = df_part.rename(columns=tmp)
# rename using the supplied mapping
df_part = df_part.rename(columns=remap)
# iterate through columns in this df partition
for col in df_part.columns:
# drop anything not in our expected list
if col not in must_haves:
df_part = df_part.drop(col, axis=1)
continue
# fixes datetime error found by Ty Mckercher and fixed by Paul Mahler
if df_part[col].dtype == 'object' and col in ['pickup_datetime', 'dropoff_datetime']:
df_part[col] = df_part[col].astype('datetime64[ms]')
continue
# if column was read as a string, recast as float
if df_part[col].dtype == 'object':
df_part[col] = df_part[col].astype('float32')
else:
# downcast from 64bit to 32bit types
# Tesla T4 are faster on 32bit ops
if 'int' in str(df_part[col].dtype):
df_part[col] = df_part[col].astype('int32')
if 'float' in str(df_part[col].dtype):
df_part[col] = df_part[col].astype('float32')
df_part[col] = df_part[col].fillna(-1)
return df_part
def coalesce_taxi_data(fraction, random_state):
base_path = 'gcs://anaconda-public-data/nyc-taxi/csv'
# list of column names that need to be re-mapped
remap = {}
remap['tpep_pickup_datetime'] = 'pickup_datetime'
remap['tpep_dropoff_datetime'] = 'dropoff_datetime'
remap['ratecodeid'] = 'rate_code'
#create a list of columns & dtypes the df must have
must_haves = {
'pickup_datetime': 'datetime64[ms]',
'dropoff_datetime': 'datetime64[ms]',
'passenger_count': 'int32',
'trip_distance': 'float32',
'pickup_longitude': 'float32',
'pickup_latitude': 'float32',
'rate_code': 'int32',
'dropoff_longitude': 'float32',
'dropoff_latitude': 'float32',
'fare_amount': 'float32'
}
# apply a list of filter conditions to throw out records with missing or outlier values
query_frags = [
'fare_amount > 0 and fare_amount < 500',
'passenger_count > 0 and passenger_count < 6',
'pickup_longitude > -75 and pickup_longitude < -73',
'dropoff_longitude > -75 and dropoff_longitude < -73',
'pickup_latitude > 40 and pickup_latitude < 42',
'dropoff_latitude > 40 and dropoff_latitude < 42'
]
valid_months_2016 = [str(x).rjust(2, '0') for x in range(1, 7)]
valid_files_2016 = [f'{base_path}/2016/yellow_tripdata_2016-{month}.csv' for month in valid_months_2016]
df_2014_fractional = dask_cudf.read_csv(f'{base_path}/2014/yellow_*.csv', chunksize=25e6).sample(
frac=fraction, random_state=random_state)
df_2014_fractional = clean(df_2014_fractional, remap, must_haves)
df_2015_fractional = dask_cudf.read_csv(f'{base_path}/2015/yellow_*.csv', chunksize=25e6).sample(
frac=fraction, random_state=random_state)
df_2015_fractional = clean(df_2015_fractional, remap, must_haves)
df_2016_fractional = dask_cudf.read_csv(valid_files_2016, chunksize=25e6).sample(
frac=fraction, random_state=random_state)
df_2016_fractional = clean(df_2016_fractional, remap, must_haves)
df_taxi = dask.dataframe.multi.concat([df_2014_fractional, df_2015_fractional, df_2016_fractional])
df_taxi = df_taxi.query(' and '.join(query_frags))
return df_taxidef taxi_csv_data_loader(client, response_dtype=np.float32, fraction=1.0, random_state=0):
response_id = 'fare_amount'
workers = client.has_what().keys()
km_fields = ['passenger_count', 'trip_distance', 'pickup_longitude', 'pickup_latitude', 'rate_code',
'dropoff_longitude', 'dropoff_latitude', 'fare_amount']
taxi_df = coalesce_taxi_data(fraction=fraction, random_state=random_state)
taxi_df = taxi_df[km_fields]
with dask.annotate(workers=set(workers)):
taxi_df = client.persist(collections=taxi_df)
X = taxi_df[taxi_df.columns.difference([response_id])].astype(np.float32)
y = taxi_df[response_id].astype(response_dtype)
wait(taxi_df)
return taxi_df, X, y
def taxi_parquet_data_loader(client, response_dtype=np.float32, fraction=1.0, random_state=0):
# list of column names that need to be re-mapped
remap = {}
remap['tpep_pickup_datetime'] = 'pickup_datetime'
remap['tpep_dropoff_datetime'] = 'dropoff_datetime'
remap['ratecodeid'] = 'rate_code'
#create a list of columns & dtypes the df must have
must_haves = {
'pickup_datetime': 'datetime64[ms]',
'dropoff_datetime': 'datetime64[ms]',
'passenger_count': 'int32',
'trip_distance': 'float32',
'pickup_longitude': 'float32',
'pickup_latitude': 'float32',
'rate_code': 'int32',
'dropoff_longitude': 'float32',
'dropoff_latitude': 'float32',
'fare_amount': 'float32'
}
# apply a list of filter conditions to throw out records with missing or outlier values
query_frags = [
'fare_amount > 0 and fare_amount < 500',
'passenger_count > 0 and passenger_count < 6',
'pickup_longitude > -75 and pickup_longitude < -73',
'dropoff_longitude > -75 and dropoff_longitude < -73',
'pickup_latitude > 40 and pickup_latitude < 42',
'dropoff_latitude > 40 and dropoff_latitude < 42'
]
workers = client.has_what().keys()
taxi_parquet_path = "gs://anaconda-public-data/nyc-taxi/nyc.parquet"
response_id = 'fare_amount'
fields = ['passenger_count', 'trip_distance', 'pickup_longitude', 'pickup_latitude', 'rate_code',
'dropoff_longitude', 'dropoff_latitude', 'fare_amount']
taxi_df = dask_cudf.read_parquet(taxi_parquet_path, npartitions=len(workers))
taxi_df = clean(taxi_df, remap, must_haves)
taxi_df = taxi_df.query(' and '.join(query_frags))
taxi_df = taxi_df[fields]
with dask.annotate(workers=set(workers)):
taxi_df = client.persist(collections=taxi_df)
wait(taxi_df)
X = taxi_df[taxi_df.columns.difference([response_id])].astype(np.float32)
y = taxi_df[response_id].astype(response_dtype)
return taxi_df, X, ydef record_elapsed_timings_to_df(df, timings, record_template, type, columns, write_to=None):
records = [dict(record_template, **{"sample_index": i,
"elapsed": elapsed,
"type": type})
for i, elapsed in enumerate(timings)]
df = cudf.concat([df, cudf.DataFrame(records)], ignore_index=True)
if (write_to):
df.to_csv(write_to, columns=columns)
return df
def collect_load_time_samples(load_func, count, return_final_sample=True, verbose=False):
timings = []
for m in tqdm(range(count)):
with SimpleTimer() as timer:
df, X, y = load_func()
timings.append(timer.elapsed)
if (return_final_sample):
return df, X, y, timings
return None, None, None, timings
def collect_func_time_samples(func, count, verbose=False):
timings = []
for k in tqdm(range(count)):
with SimpleTimer() as timer:
func()
timings.append(timer.elapsed)
return timings
def sweep_fit_func(model, func_id, require_compute, X, y, xy_fit, count):
_fit_func_attr = getattr(model, func_id)
if (require_compute):
if (xy_fit):
fit_func = partial(lambda X, y: _fit_func_attr(X, y).compute(), X, y)
else:
fit_func = partial(lambda X: _fit_func_attr(X).compute(), X)
else:
if (xy_fit):
fit_func = partial(_fit_func_attr, X, y)
else:
fit_func = partial(_fit_func_attr, X)
return collect_func_time_samples(func=fit_func, count=count)
def sweep_predict_func(model, func_id, require_compute, X, count):
_predict_func_attr = getattr(model, func_id)
predict_func = partial(lambda X: _predict_func_attr(X).compute(), X)
return collect_func_time_samples(func=predict_func, count=count)
def performance_sweep(client, model, data_loader, hardware_type, worker_counts=[1], samples=1, load_samples=1, max_data_frac=1.0,
predict_frac=0.05, scaling_type='weak', xy_fit=True, fit_requires_compute=False, update_workers_in_kwargs=True,
response_dtype=np.float32, out_path='./perf_sweep.csv', append_to_existing=False, model_name=None,
fit_func_id="fit", predict_func_id="predict", scaling_denom=None, model_args={}, model_kwargs={}):
"""
Primary performance sweep entrypoint.
Parameters
------------
client: DASK client associated with the cluster we're interesting in collecting performance data for.
model: Model object on which to gather performance data. This will be created and destroyed,
once for each element of 'worker_counts'
data_loader: arbitrary data loading function that will be called to load the appropriate testing data.
Function that is responsible for loading and returning the data to be used for a given performance run. Function
signature must accept (client, fraction, and random_state). Client should be used to distribute data, and loaders
should utilize fraction and random_state with dask's dataframe.sample method to allow for control of how much data
is loaded.
When called, its return value should be of the form: df, X, y, where df is the full dask_cudf dataframe, X is a
dask_cudf dataframe which contains all explanatory variables that will be passed to the 'fit' function, and y is a
dask_cudf series or dataframe that contains response variables which should be passed to fit/predict as fit(X, y)
hardware_type: indicates the core hardware the current sweep is running on. ex. 'T4', 'V100', 'A100'
worker_counts: List indicating the number of workers that should be swept. Ex [1, 2, 4]
worker counts must fit within the cluster associated with 'client', if the current DASK worker count is different
from what is requested on a given sweep, attempt to automatically scale the worker count. NOTE: this does not
mean we will scale the available cluster nodes, just the number of deployed worker pods.
samples: number of fit/predict samples to record per worker count
load_samples: number of times to sample data loads. This effectively times how long 'data_loader' runs.
max_data_frac: maximum fraction of data to return.
Strong scaling: each run will utilize max_data_frac data.
Weak scaling: each run will utilize (current worker count) / (max worker count) * max_data_frac data.
predict_frac: fraction of training data used to test inference
scaling_type: values can be 'weak' or 'strong' indicating the type of scaling sweep to perform.
xy_fit: indicates whether or not the model's 'fit' function is of the form (X, y), when xy_fit is False, we assume that
fit is of the form (X), as is the case with various unsupervised methods ex. KNN.
fit_requires_compute: False generally, set this to True if the model's 'fit' function requires a corresponding '.compute()'
call to execute the required work.
update_workers_in_kwargs: Some algorithms accept a 'workers' list, much like DASK, and will require their kwargs to have
workers populated. Setting this flag handles this automatically.
response_dtype: defaults to np.float32, some algorithms require another dtype, such as int32
out_path: path where performance data csv should be saved
append_to_existing: When true, append results to an existing csv, otherwise overwrite.
model_name: Override what we output as the model name
fit_func_id: Defaults to 'fit', only set this if the model has a non-standard naming.
predict_func_id: Defaults to 'predict', only set this if the model has a on-standard predict naming.
scaling_denom: (weak scaling) defaults to max(workers) if unset. Specifies the maximum worker count that weak scaling
should scale against. For example, when using 1 worker in a weak scaling sweep, the worker will attempt to
process a fraction of the total data equal to 1/scaling_denom
model_args: args that will be passed to the model's constructor
model_kwargs: keyword args that will be passed to the model's constructor
Returns
--------
"""
cols = ['n_workers', 'sample_index', 'elapsed', 'type', 'algorithm', 'scaling_type', 'data_fraction', 'hardware']
perf_df = cudf.DataFrame(columns=cols)
if (append_to_existing):
try:
perf_df = cudf.read_csv(out_path)
except:
pass
model_name = model_name if model_name else str(model)
scaling_denom = scaling_denom if (scaling_denom is not None) else max(worker_counts)
max_data_frac = min(1.0, max_data_frac)
start_msg = f"Starting {scaling_type}-scaling performance sweep for:\n"
start_msg += f" model : {model_name}\n"
start_msg += f" data loader: {data_loader}.\n"
start_msg += f"Configuration\n"
start_msg += "==========================\n"
start_msg += f"{'Worker counts':<25} : {worker_counts}\n"
start_msg += f"{'Fit/Predict samples':<25} : {samples}\n"
start_msg += f"{'Data load samples':<25} : {load_samples}\n"
start_msg += f"- {'Max data fraction':<23} : {max_data_frac}\n"
start_msg += f"{'Model fit':<25} : {'X ~ y' if xy_fit else 'X'}\n"
start_msg += f"- {'Response DType':<23} : {response_dtype}\n"
start_msg += f"{'Writing results to':<25} : {out_path}\n"
start_msg += f"- {'Method':<23} : {'overwrite' if not append_to_existing else 'append'}\n"
print(start_msg, flush=True)
for n in worker_counts:
fraction = (n / scaling_denom) * max_data_frac if scaling_type == 'weak' else max_data_frac
record_template = {"n_workers": n, "type": "predict", "algorithm": model_name,
"scaling_type": scaling_type, "data_fraction": fraction, "hardware": hardware_type}
scale_workers(client, n)
print(f"Sampling <{load_samples}> load times with {n} workers.", flush=True)
load_func = partial(data_loader, client=client, response_dtype=response_dtype, fraction=fraction, random_state=0)
df, X, y, load_timings = collect_load_time_samples(load_func=load_func, count=load_samples)
perf_df = record_elapsed_timings_to_df(df=perf_df, timings=load_timings, type='load',
record_template=record_template, columns=cols, write_to=out_path)
print(f"Finished loading <{load_samples}>, samples, to <{n}> workers with a mean time of {np.mean(load_timings)/1e9:0.4f} sec.", flush=True)
print(f"Sweeping {model_name} '{fit_func_id}' with <{n}> workers. Sampling <{samples}> times.", flush=True)
if (update_workers_in_kwargs and 'workers' in model_kwargs):
model_kwargs['workers'] = workers = list(client.has_what().keys())
m = model(*model_args, **model_kwargs)
if (fit_func_id):
fit_timings = sweep_fit_func(model=m, func_id=fit_func_id,
require_compute=fit_requires_compute,
X=X, y=y, xy_fit=xy_fit, count=samples)
perf_df = record_elapsed_timings_to_df(df=perf_df, timings=fit_timings, type='fit',
record_template=record_template, columns=cols, write_to=out_path)
print(f"Finished gathering <{samples}>, 'fit' samples using <{n}> workers, with a mean time of {np.mean(fit_timings)/1e9:0.4f} sec.",
flush=True)
else:
print(f"Skipping fit sweep, fit_func_id is None")
if (predict_func_id):
print(f"Sweeping {model_name} '{predict_func_id}' with <{n}> workers. Sampling <{samples}> times.", flush=True)
predict_timings = sweep_predict_func(model=m, func_id=predict_func_id,
require_compute=True, X=X, count=samples)
perf_df = record_elapsed_timings_to_df(df=perf_df, timings=predict_timings, type='predict',
record_template=record_template, columns=cols, write_to=out_path)
print(f"Finished gathering <{samples}>, 'predict' samples using <{n}> workers, with a mean time of {np.mean(predict_timings)/1e9:0.4f} sec.",
flush=True)
else:
print(f"Skipping inference sweep. predict_func_id is None")def simple_ci(df, fields, groupby):
gbdf = df[fields].groupby(groupby).agg(['mean', 'std', 'count'])
ci = (1.96 + gbdf['elapsed']['std'] / np.sqrt(gbdf['elapsed']['count']))
ci_df = ci.reset_index()
ci_df['ci.low'] = gbdf['elapsed'].reset_index()['mean'] - ci_df[0]
ci_df['ci.high'] = gbdf['elapsed'].reset_index()['mean'] + ci_df[0]
return ci_df
def visualize_csv_data(csv_path):
df = cudf.read_csv(csv_path)
fields = ['elapsed', 'elapsed_sec', 'type', 'n_workers', 'hardware', 'scaling_type']
groupby = ['n_workers', 'type', 'hardware', 'scaling_type']
df['elapsed_sec'] = df['elapsed']/1e9
ci_df = simple_ci(df, fields, groupby=groupby)
# Rescale to seconds
ci_df[['ci.low', 'ci.high']] = ci_df[['ci.low', 'ci.high']]/1e9
# Print confidence intervals
print(ci_df[['hardware', 'n_workers', 'type', 'ci.low', 'ci.high']][ci_df['type'] != 'load'])
sns.set_theme(style="whitegrid")
sns.set(rc={'figure.figsize':(20, 10)}, font_scale=2)
# Boxplots for elapsed time at each worker count.
plot_df = df[fields][df[fields].type != 'load'].to_pandas()
ax = sns.catplot(data=plot_df, x="n_workers", y="elapsed_sec",
col="type", row="scaling_type", hue="hardware", kind="box",
height=8, order=None)# Uncomment to test with Taxi Dataset
preload_data = False
append_to_existing = True
samples = 5
load_samples = 1
worker_counts = [8]
scaling_denom = 8
hardware_type = None
max_data_frac = 1.0
scale_type = 'weak' # weak | strong
out_prefix = 'taxi_medium'
if (not preload_data):
data_loader = taxi_parquet_data_loader
else:
data = taxi_parquet_data_loader(client, fraction=max_data_frac)
data_loader = lambda client, response_dtype, fraction, random_state: data
if (not hardware_type):
raise RuntimeError("Please specify the hardware type for this run! ex. (T4, V100, A100)")
sweep_kwargs = {
'append_to_existing': append_to_existing,
'samples': samples,
'load_samples': load_samples,
'worker_counts': worker_counts,
'scaling_denom': scaling_denom,
'hardware_type': hardware_type,
'data_loader': data_loader,
'max_data_frac': max_data_frac,
'scaling_type': scale_type
}taxi_parquet_path = ["gs://anaconda-public-data/nyc-taxi/nyc.parquet"]
estimated_rows = estimate_df_rows(client, files=taxi_parquet_path, testpct=0.0001)
print(estimated_rows)# Uncomment to sweep with the large Taxi Dataset
preload_data = True
append_to_existing = True
samples = 5
load_samples = 1
worker_counts = [8]
scaling_denom = 8
hardware_type = None
data_loader = taxi_csv_data_loader
max_data_frac = 1.0
scale_type = 'weak'
out_prefix = 'taxi_large'
if (not preload_data):
data_loader = taxi_csv_data_loader
else:
data = taxi_csv_data_loader(client, fraction=max_data_frac)
data_loader = lambda client, response_dtype, fraction, random_state: data
if (not hardware_type):
raise RuntimeError("Please specify the hardware type for this run! ex. (T4, V100, A100)")
sweep_kwargs = {
'append_to_existing': append_to_existing,
'samples': samples,
'load_samples': load_samples,
'worker_counts': worker_counts,
'scaling_denom': scaling_denom,
'hardware_type': hardware_type,
'data_loader': data_loader,
'max_data_frac': max_data_frac,
'scaling_type': scale_type
}remap = {}
remap['tpep_pickup_datetime'] = 'pickup_datetime'
remap['tpep_dropoff_datetime'] = 'dropoff_datetime'
remap['ratecodeid'] = 'rate_code'
#create a list of columns & dtypes the df must have
must_haves = {
'pickup_datetime': 'datetime64[ms]',
'dropoff_datetime': 'datetime64[ms]',
'passenger_count': 'int32',
'trip_distance': 'float32',
'pickup_longitude': 'float32',
'pickup_latitude': 'float32',
'rate_code': 'int32',
'dropoff_longitude': 'float32',
'dropoff_latitude': 'float32',
'fare_amount': 'float32'
}
# apply a list of filter conditions to throw out records with missing or outlier values
query_frags = [
'fare_amount > 0 and fare_amount < 500',
'passenger_count > 0 and passenger_count < 6',
'pickup_longitude > -75 and pickup_longitude < -73',
'dropoff_longitude > -75 and dropoff_longitude < -73',
'pickup_latitude > 40 and pickup_latitude < 42',
'dropoff_latitude > 40 and dropoff_latitude < 42'
]
workers = client.has_what().keys()base_path = 'gcs://anaconda-public-data/nyc-taxi/csv'
with SimpleTimer() as timer_csv:
df_csv_2014 = dask_cudf.read_csv(f'{base_path}/2014/yellow_*.csv', chunksize=25e6)
df_csv_2014 = clean(df_csv_2014, remap, must_haves)
df_csv_2014 = df_csv_2014.query(' and '.join(query_frags))
with dask.annotate(workers=set(workers)):
df_csv_2014 = client.persist(collections=df_csv_2014)
wait(df_csv_2014)
print(df_csv_2014.columns)
rows_csv = df_csv_2014.iloc[:,0].shape[0].compute()
print(f"CSV load took {timer_csv.elapsed/1e9} sec. For {rows_csv} rows of data => {rows_csv/(timer_csv.elapsed/1e9)} rows/sec")client.cancel(df_csv_2014)with SimpleTimer() as timer_parquet:
df_parquet = dask_cudf.read_parquet(f'gs://anaconda-public-data/nyc-taxi/nyc.parquet', chunksize=25e6)
df_parquet = clean(df_parquet, remap, must_haves)
df_parquet = df_parquet.query(' and '.join(query_frags))
with dask.annotate(workers=set(workers)):
df_parquet = client.persist(collections=df_parquet)
wait(df_parquet)
print(df_parquet.columns)
rows_parquet = df_parquet.iloc[:,0].shape[0].compute()
print(f"Parquet load took {timer_parquet.elapsed/1e9} sec. For {rows_parquet} rows of data => {rows_parquet/(timer_parquet.elapsed/1e9)} rows/sec")client.cancel(df_parquet)speedup = (rows_parquet/(timer_parquet.elapsed/1e9))/(rows_csv/(timer_csv.elapsed/1e9))
print(speedup)from cuml.dask.ensemble import RandomForestRegressor
rf_kwargs = {
"workers": client.has_what().keys(),
"n_estimators": 10,
"max_depth": 12
}
rf_csv_path = f"./{out_prefix}_random_forest_regression.csv"
performance_sweep(client=client, model=RandomForestRegressor,
**sweep_kwargs,
out_path=rf_csv_path,
response_dtype=np.int32,
model_kwargs=rf_kwargs)rf_csv_path = f"./{out_prefix}_random_forest_regression.csv"
visualize_csv_data(rf_csv_path)from cuml.dask.cluster import KMeans
kmeans_kwargs = {
"client": client,
"n_clusters": 12,
"max_iter": 371,
"tol": 1e-5,
"oversampling_factor": 3,
"max_samples_per_batch": 32768/2,
"verbose": False,
"init": 'random'
}
kmeans_csv_path = f'./{out_prefix}_kmeans.csv'
performance_sweep(client=client, model=KMeans,
**sweep_kwargs,
out_path=kmeans_csv_path,
xy_fit=False,
model_kwargs=kmeans_kwargs)visualize_csv_data(kmeans_csv_path)from cuml.dask.neighbors import NearestNeighbors
nn_kwargs = {}
nn_csv_path = f'./{out_prefix}_nn.csv'
performance_sweep(client=client, model=NearestNeighbors,
**sweep_kwargs,
out_path=nn_csv_path,
xy_fit=False,
predict_func_id='get_neighbors',
model_kwargs=nn_kwargs)nn_csv_path = f'./{out_prefix}_nn.csv'
visualize_csv_data(nn_csv_path)from cuml.dask.decomposition import PCA
pca_kwargs = {
"client": client,
"n_components": 5,
"whiten": False
}
pca_csv_path = f'./{out_prefix}_pca.csv'
performance_sweep(client=client, model=PCA,
**sweep_kwargs,
out_path=pca_csv_path,
xy_fit=False,
fit_requires_compute=True,
fit_func_id="fit_transform",
predict_func_id=None, # PCA has no 'predict' method.
model_kwargs=pca_kwargs)visualize_csv_data(pca_csv_path)from cuml.dask.decomposition import TruncatedSVD
tsvd_kwargs = {
"client": client,
"n_components": 5
}
tsvd_csv_path = f'./{out_prefix}_tsvd.csv'
performance_sweep(client=client, model=TruncatedSVD,
**sweep_kwargs,
out_path=tsvd_csv_path,
xy_fit=False,
fit_requires_compute=True,
fit_func_id="fit_transform",
predict_func_id=None,
model_kwargs=tsvd_kwargs)visualize_csv_data(tsvd_csv_path)from cuml.dask.linear_model import LinearRegression
lr_kwargs = {
"client": client,
"algorithm": "eig"
}
lr_csv_path = f'./{out_prefix}_linear_regression.csv'
performance_sweep(client=client, model=LinearRegression,
**sweep_kwargs,
out_path=lr_csv_path,
model_kwargs=lr_kwargs)visualize_csv_data(lr_csv_path)from cuml.dask.linear_model import Ridge as RidgeRegression
ridge_kwargs = {
"client": client,
"solver": "eig"
}
ridge_csv_path = f'./{out_prefix}_ridge_regression.csv'
performance_sweep(client=client, model=RidgeRegression,
**sweep_kwargs,
out_path=ridge_csv_path,
model_kwargs=ridge_kwargs)visualize_csv_data(ridge_csv_path)from cuml.dask.linear_model import Lasso as LassoRegression
lasso_kwargs = {
"client": client
}
lasso_csv_path = f'./{out_prefix}_lasso_regression.csv'
performance_sweep(client=client, model=LassoRegression,
**sweep_kwargs,
out_path=lasso_csv_path,
model_kwargs=lasso_kwargs)visualize_csv_data(lasso_csv_path)from cuml.dask.linear_model import ElasticNet as ElasticNetRegression
elastic_kwargs = {
"client": client,
}
enr_csv_path = f'./{out_prefix}_elastic_regression.csv'
performance_sweep(client=client, model=ElasticNetRegression,
**sweep_kwargs,
out_path=enr_csv_path,
model_kwargs=elastic_kwargs)visualize_csv_data(enr_csv_path)from cuml.dask.solvers import CD
cd_kwargs = {
}
cd_csv_path = f'./{out_prefix}_mutli_gpu_linear_regression.csv'
performance_sweep(client=client, model=CD,
**sweep_kwargs,
out_path=cd_csv_path,
model_kwargs=cd_kwargs)visualize_csv_data(cd_csv_path)import xgboost as xgb
xg_args = [client]
xg_kwargs = {
'params': {
'tree_method': 'gpu_hist',
},
'num_boost_round': 100
}
xgb_csv_path = f'./{out_prefix}_xgb.csv'
class XGBProxy():
"""
Create a simple API wrapper around XGBoost so that it supports the fit/predict workflow.
Parameters
-------------
data_loader: data loader object intended to be used by the performance sweep.
"""
def __init__(self, data_loader):
self.args = []
self.kwargs = {}
self.data_loader = data_loader
self.trained_model = None
def loader(self, client, response_dtype, fraction, random_state):
"""
Wrap the data loader method so that it creates a DMatrix from the returned data.
"""
df, X, y = self.data_loader(client, response_dtype, fraction, random_state)
dmatrix = xgb.dask.DaskDMatrix(client, X, y)
return dmatrix, dmatrix, dmatrix
def __call__(self, *args, **kwargs):
"""
Acts as a pseudo init function which initializes our model args.
"""
self.args = args
self.kwargs = kwargs
return self
def fit(self, X):
"""
Wrap dask.train, and store the model on our proxy object.
"""
if (self.trained_model):
del self.trained_model
self.trained_model = xgb.dask.train(*self.args,
dtrain=X,
evals=[(X, 'train')],
**self.kwargs)
return self
def predict(self, X):
assert(self.trained_model)
return xgb.dask.predict(*self.args, self.trained_model, X)
xgb_proxy = XGBProxy(data_loader)
performance_sweep(client=client, model=xgb_proxy, data_loader=xgb_proxy.loader, hardware_type=hardware_type,
worker_counts=worker_counts,
samples=samples,
load_samples=load_samples,
max_data_frac=max_data_frac,
scaling_type=scale_type,
out_path=xgb_csv_path,
append_to_existing=append_to_existing,
update_workers_in_kwargs=False,
xy_fit=False,
scaling_denom = scaling_denom,
model_args=xg_args,
model_kwargs=xg_kwargs)visualize_csv_data(xgb_csv_path) | 0 |
rapidsai_public_repos/cloud-ml-examples/dask | rapidsai_public_repos/cloud-ml-examples/dask/kubernetes/Dockerfile | FROM rapidsai/rapidsai-core:22.06-cuda11.5-runtime-ubuntu20.04-py3.9
# Install required package for notebook and cluster control
RUN mamba install -n rapids -c conda-forge --freeze-installed -y kubernetes google-cloud-sdk gcsfs seaborn dask-kubernetes
# Install gke-gcloud-auth-plugin, see https://cloud.google.com/blog/products/containers-kubernetes/kubectl-auth-changes-in-gke
RUN gcloud components install gke-gcloud-auth-plugin
RUN git clone --depth 1 https://github.com/rapidsai/cloud-ml-examples.git /rapids/notebooks/cloud-ml-examples
| 0 |
rapidsai_public_repos/cloud-ml-examples/dask/kubernetes | rapidsai_public_repos/cloud-ml-examples/dask/kubernetes/specs/sched-spec.yaml | apiVersion: v1
kind: Pod
metadata:
name: dask-scheduler
labels:
cluster_type: dask
dask_type: scheduler
spec:
restartPolicy: Never
containers:
- image: rapidsai/rapidsai-core:22.06-cuda11.5-runtime-ubuntu20.04-py3.9
imagePullPolicy: IfNotPresent
env:
- name: DISABLE_JUPYTER
value: "true"
- name: EXTRA_PIP_PACKAGES
value: "gcsfs"
args: [ dask-scheduler ]
name: dask-scheduler
resources:
limits:
cpu: "2"
memory: 3G
requests:
cpu: "2"
memory: 3G
| 0 |
rapidsai_public_repos/cloud-ml-examples/dask/kubernetes | rapidsai_public_repos/cloud-ml-examples/dask/kubernetes/specs/worker-spec.yaml | apiVersion: v1
kind: Pod
metadata:
labels:
cluster_type: dask
dask_type: GPU_worker
spec:
restartPolicy: Never
containers:
- image: rapidsai/rapidsai-core:22.06-cuda11.5-runtime-ubuntu20.04-py3.9
imagePullPolicy: IfNotPresent
env:
- name: DISABLE_JUPYTER
value: "true"
- name: EXTRA_PIP_PACKAGES
value: "gcsfs"
args: [ dask-cuda-worker, $(DASK_SCHEDULER_ADDRESS), --rmm-managed-memory ]
name: dask-cuda-worker
resources:
limits:
cpu: "2"
memory: 3G
nvidia.com/gpu: 1
requests:
cpu: "2"
memory: 3G
nvidia.com/gpu: 1
| 0 |
rapidsai_public_repos/cloud-ml-examples | rapidsai_public_repos/cloud-ml-examples/databricks/README.md | # RAPIDS on Databricks
This directory contains sample notebooks for running RAPIDS on Databricks.
The `rapids_intro.ipynb` notebook has been tested with the latest RAPIDS version (0.19) by building a custom container, and contains basic examples to get started with cuDF and cuML.
The `rapids_airline_hyperopt.ipynb` example walks through the optimization of a random forest model using cuML and hyperopt. It includes init scripts to install an earlier version of RAPIDS (0.13) on DataBricks ML Runtime.
## 1. Use a custome image on Databricks
## Build the RAPIDS container
```console
$ docker build --tag <username>/rapids_databricks:latest --build-arg RAPIDS_IMAGE=rapidsai/rapidsai-core:22.06-cuda11.0-base-ubuntu18.04-py3.8 ./docker
```
Push this image to a Docker registry (DockerHub, Amazon ECR or Azure ACR).
## Configure and create a cluster
* Create your cluster:
1. Select a standard Databricks runtime. In this example 8.2 version, since we're using a container with CUDA 11.
* This needs to be a Databricks runtime version that supports Databricks Container Services.
2. Select "Use your own Docker container".
3. In the Docker Image URL field, enter the image that you created above.
4. Select a GPU enabled worker and driver type.
* **Note** Selected GPU must be Pascal generation or greater.
5. Create and launch your cluster.
## Launching the notebook example
1. Upload the `rapids_intro.ipynb` notebook to your workspace.
2. Execute the cells to import cuDF and cuML, and walk through simple examples on the GPU.
## 2. Use an init script on Databricks
**The example below has been tested with an earlier version of RAPIDS (0.13). To use the latest version of RAPIDS, follow the steps mentioned above.**
### Upload RAPIDS 0.13 Init Script to DBFS
* Copy `src/rapids_install_cuml0.13_cuda10.0_ubuntu16.04.sh` onto your Databricks dbfs file system.
* This will become the base init script that is run at cluster start up.
* Example:
```shell script
$ dbfs configure
... configure your dbfs client for your account ...
$ dbfs cp src/rapids_install_cuml0.13_cuda10.0_ubuntu16.04.sh dbfs:/databricks/init_scripts/
```
### Create and Configure a Cluster
* Create your cluster:
1. Select a GPU enabled Databricks runtime. Ex: 6.6 ML
* Currently 'Use your own Docker container' is not available for ML instances.
2. Select a GPU enabled worker and driver type
* **Note** Selected GPU must be Pascal generation or greater. p2.X is not supported.
* Recommended: `g4dn.xxxx` (NVIDIA T4) or `p3.xxxx` (NVIDIA V100) for AWS users
3. Select `Advanced` -> `init_scripts`
* Add an init scripts with the location `dbfs:/databricks/init_scripts/rapids_instal_cuml0.13_cuda10.0_ubuntu16.04.sh`
* Launch your cluster
* At this point, you should have RAPIDS 0.13 installed in the databricks-ml-gpu conda environment, and can import cudf/cuml modules.
### Launching the notebook
1. Upload the `rapids_airline_hyperopt.ipynb` notebook to your workspace.
2. Uncomment the "data download" cell and configure it to point to a path of your choice for data download. By default, it will use a smaller (200k row) dataset. This executes fast but doesn't demonstrate the full speedups possible with larger datasets.
3. Execute all of the cells to launch your hyperopt job.
4. Optionally, check out stats in the runs page and Experiment UI.
## More on Integrating Databricks Jobs with MLFlow and RAPIDS
You can find more detail in [this blog post on MLFlow + RAPIDS](https://medium.com/rapids-ai/managing-and-deploying-high-performance-machine-learning-models-on-gpus-with-rapids-and-mlflow-753b6fcaf75a).
| 0 |
rapidsai_public_repos/cloud-ml-examples/databricks | rapidsai_public_repos/cloud-ml-examples/databricks/notebooks/rapids_airline_hyperopt.ipynb | import warnings
warnings.filterwarnings("ignore", category=DeprecationWarning)
warnings.filterwarnings("ignore", category=FutureWarning)
import cudf
import cuml
import mlflow
import hyperopt
import numpy as np
import pandas as pd
import mlflow.sklearn
from mlflow.tracking.client import MlflowClient
from hyperopt import fmin, tpe, hp, Trials, STATUS_OK
import cuml.ensemble
import cuml.metrics
import cuml.preprocessing.model_selection
import sklearn.ensemble
import sklearn.metrics
import sklearn.model_selection# Small utility that times a block of code and prints how long it took to execute
from contextlib import contextmanager
import time
@contextmanager
def timed(name):
t0 = time.time()
yield
t1 = time.time()
print("..%-24s: %8.4f" % (name, t1 - t0))MAX_EVALS = 20
MAX_PARALLEL = 2# # Read above instructions - RUN ONLY ONCE
# from urllib.request import urlretrieve
# import os
# file_name = 'airline_small.parquet' # NOTE: Change to airline_20000000.parquet to use a larger dataset
# data_dir = "/_dbfs_p8ath/" # NOTE: Change to DBFS path where you want to save the file
# INPUT_FILE = os.path.join(data_dir, file_name)
# if os.path.isfile(INPUT_FILE):
# print(f" > File already exists. Ready to load at {INPUT_FILE}")
# else:
# # Ensure folder exists
# os.makedirs(data_dir, exist_ok=True)
# url = "https://rapidsai-cloud-ml-sample-data.s3-us-west-2.amazonaws.com/" + file_name
# urlretrieve(url= url,filename=INPUT_FILE)
# print("Completed!")df = cudf.read_parquet(INPUT_FILE)
print("Data shape: ", df.shape)
df.head()def train_cpu(params, test_set_frac=0.2, registered_model_name=None):
"""
Train scikit-learn model on the data, and calculate the accuracy on the model.
This method will be passed to `hyperopt.fmin()`.
Params:
params - dict; The range of the HPO space for different parameters (max_depth, max_features, n_estimators)
in that order.
test_set_frac - float; Value between (0,1) for the size of the test set to be used for validation split
registered_model_name - string; Name under which the best model should be registered with MLFlow.
Returns:
dict with fields 'loss' (scalar loss) and 'status' (success/failure status of run)
"""
max_depth, max_features, n_estimators = params
with timed("load"):
df = pd.read_parquet(INPUT_FILE)
with timed("etl"):
X = df.drop(["ArrDelayBinary"], axis=1)
y = df["ArrDelayBinary"].astype('int32')
X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(X, y,
test_size=test_set_frac,
random_state=123)
with timed("fit"):
mod = sklearn.ensemble.RandomForestClassifier(max_depth=max_depth,
max_features=max_features,
n_estimators=n_estimators,
n_jobs=-1) # Use all available CPUs
mod.fit(X_train, y_train)
mlflow.sklearn.log_model(mod, "RF_model_cpu_large_",
registered_model_name=registered_model_name)
with timed("predict"):
if test_set_frac > 0.0:
preds = mod.predict(X_test)
acc = sklearn.metrics.accuracy_score(y_test, preds)
mlflow.log_metric("accuracy", acc)
else:
acc = np.nan
# Returning -1 * acc because fmin minimizes the "loss" and we want to maximize accuracy.
return {'loss': -acc, 'status': STATUS_OK}# Example run with a sample parameter value
with timed("sample train skl"):
result = train_cpu((8, 1.0, 100))def train_rapids(params, test_set_frac=0.2, registered_model_name=None):
"""
Train RAPIDS cuml model on the data, and calculate the accuracy on the model.
This method will be passed to `hyperopt.fmin()`.
Params and Return values same as train_cpu
"""
max_depth, max_features, n_estimators = params
# Read using cudf
with timed("read_raw"):
df = cudf.read_parquet(INPUT_FILE)
# Converting to dtypes expected by cuml model
with timed("etl"):
X = df.drop(["ArrDelayBinary"], axis=1)
y = df["ArrDelayBinary"].astype('int32')
# Splitting the data into 80/20 for training and validation
X_train, X_test, y_train, y_test = cuml.preprocessing.model_selection.train_test_split(
X, y,
test_size=test_set_frac,
random_state=123)
with timed("fit"):
n_bins = 16
mod = cuml.ensemble.RandomForestClassifier(max_depth=max_depth,
max_features=max_features,
n_bins=n_bins,
n_estimators=n_estimators)
mod.fit(X_train, y_train)
mlflow.sklearn.log_model(mod, "RF_model_GPU_",
registered_model_name=registered_model_name)
with timed("predict"):
if test_set_frac > 0.0:
preds = mod.predict(X_test)
acc = cuml.metrics.accuracy_score(y_test, preds)
mlflow.log_metric("accuracy", acc)
else:
acc = np.nan
# Returning -1 * acc because fmin minimizes the "loss" and we want to maximize accuracy.
return {'loss': -acc, 'status': STATUS_OK}# Example run with a sample parameter value
with timed("sample train rapids"):
result = train_rapids((8, 1.0, 100))# Shared search parameters
from hyperopt.pyll import scope
search_space = [
scope.int(hp.quniform('max_depth', 5, 15, 1)),
hp.uniform('max_features', 0., 1.0),
scope.int(hp.quniform('n_estimators', 100, 500, 100))
]
algo = tpe.suggest
spark.conf.set('spark.task.maxFailures', '1')spark_trials = hyperopt.SparkTrials(parallelism=MAX_PARALLEL)
mlflow.end_run() # Close out any run in progress
with mlflow.start_run() as run:
mlflow.set_tag("mlflow.runName", "CPU_run")
best = fmin(
fn=train_cpu,
space=search_space,
algo=algo,
trials = spark_trials,
max_evals=MAX_EVALS)
mlflow.set_tag("best params", str(best))
# Re-fit the best model on ALL of the data (no test set)
print(best)
train_cpu((int(best["max_depth"]), best["max_features"], int(best["n_estimators"])),
test_set_frac=0.0, registered_model_name="MLFlow_Airline_CPU_large_")
mlflow.end_run()# SparkTrials object will automatically log runs to MLFlow in DataBricks
spark_trials = hyperopt.SparkTrials(parallelism=MAX_PARALLEL)
mlflow.end_run() # Close out any run in progress
with mlflow.start_run() as run:
mlflow.set_tag("mlflow.runName", "GPU_run")
best = fmin(fn=train_rapids,
space=search_space,
algo=algo,
trials = spark_trials,
max_evals=MAX_EVALS)
mlflow.set_tag("best params", str(best))
# Re-fit the best model on ALL of the data (no test set)
train_rapids((int(best["max_depth"]), best["max_features"], int(best["n_estimators"])),
test_set_frac=0.0, registered_model_name="MLFlow_Airline_RAPIDS")
mlflow.end_run() | 0 |
rapidsai_public_repos/cloud-ml-examples/databricks | rapidsai_public_repos/cloud-ml-examples/databricks/notebooks/rapids_intro.ipynb | import cudf
import io, requests
# Download CSV file from GitHub
url="https://github.com/plotly/datasets/raw/master/tips.csv"
content = requests.get(url).content.decode('utf-8')
# Read CSV from memory
tips_df = cudf.read_csv(io.StringIO(content))
tips_df['tip_percentage'] = tips_df['tip']/tips_df['total_bill']*100
# Display average tip by dining party size
print(tips_df.groupby('size').tip_percentage.mean())from cuml import make_regression, train_test_split
from cuml.linear_model import LinearRegression as cuLinearRegression
from cuml.metrics.regression import r2_score
from sklearn.linear_model import LinearRegression as skLinearRegression
# Define parameters
n_samples = 2**20 #If you are running on a GPU with less than 16GB RAM, please change to 2**19 or you could run out of memory
n_features = 399
random_state = 23# Generate data
%%time
X, y = make_regression(n_samples=n_samples, n_features=n_features, random_state=random_state)
X = cudf.DataFrame(X)
y = cudf.DataFrame(y)[0]
X_cudf, X_cudf_test, y_cudf, y_cudf_test = train_test_split(X, y, test_size = 0.2, random_state=random_state)# Copy dataset from GPU memory to host memory (CPU)
# This is done to later compare CPU and GPU results
X_train = X_cudf.to_pandas()
X_test = X_cudf_test.to_pandas()
y_train = y_cudf.to_pandas()
y_test = y_cudf_test.to_pandas()%%time
ols_sk = skLinearRegression(fit_intercept=True,
normalize=True,
n_jobs=-1)
ols_sk.fit(X_train, y_train)%%time
predict_sk = ols_sk.predict(X_test)%%time
r2_score_sk = r2_score(y_cudf_test, predict_sk)%%time
ols_cuml = cuLinearRegression(fit_intercept=True,
normalize=True,
algorithm='eig')
ols_cuml.fit(X_cudf, y_cudf)%%time
predict_cuml = ols_cuml.predict(X_cudf_test)%%time
r2_score_cuml = r2_score(y_cudf_test, predict_cuml)print("R^2 score (SKL): %s" % r2_score_sk)
print("R^2 score (cuML): %s" % r2_score_cuml) | 0 |
rapidsai_public_repos/cloud-ml-examples/databricks | rapidsai_public_repos/cloud-ml-examples/databricks/docker/Dockerfile | ARG RAPIDS_IMAGE
FROM $RAPIDS_IMAGE as rapids
RUN conda list -n rapids --explicit > /rapids/rapids-spec.txt
FROM databricksruntime/gpu-conda:cuda11
COPY --from=rapids /rapids/rapids-spec.txt /tmp/spec.txt
RUN conda create --name rapids --file /tmp/spec.txt && \
rm -f /tmp/spec.txt
# Set an environment variable used by Databricks to decide which conda environment to activate by default.
ENV DEFAULT_DATABRICKS_ROOT_CONDA_ENV=rapids
| 0 |
rapidsai_public_repos/cloud-ml-examples/databricks | rapidsai_public_repos/cloud-ml-examples/databricks/src/rapids_install_cuml0.13_cuda10.0_ubuntu16.04.sh | #!/usr/bin/env bash
set -x
set -e
/databricks/python/bin/python -V
. /databricks/conda/etc/profile.d/conda.sh
conda activate /databricks/python
INSTALL_FILE="/opt/rapids_initialized.log"
if [[ -f "$INSTALL_FILE" ]]; then
TEST=$(cat "$INSTALL_FILE")
if (( $TEST == 1 )); then
echo "Node was previously configured. Exiting."
exit 0
fi
fi
cat > rapids0.13_cuda10.0_ubuntu16.04.yml <<EOF
name: databricks-ml-gpu
channels:
- rapidsai
- nvidia
- conda-forge
dependencies:
- _libgcc_mutex=0.1=conda_forge
- _openmp_mutex=4.5=1_llvm
- arrow-cpp=0.15.0=py37h090bef1_2
- bokeh=2.0.1=py37hc8dfbb8_0
- boost-cpp=1.70.0=h8e57a91_2
- brotli=1.0.7=he1b5a44_1002
- bzip2=1.0.8=h516909a_2
- c-ares=1.15.0=h516909a_1001
- ca-certificates=2020.4.5.1=hecc5488_0
- certifi=2020.4.5.1=py37hc8dfbb8_0
- click=7.1.2=pyh9f0ad1d_0
- cloudpickle=1.4.1=py_0
- cudatoolkit=10.0.130=0
- cudf=0.13.0=py37_0
- cudnn=7.6.0=cuda10.0_0
- cuml=0.13.0=cuda10.0_py37_0
- cupy=7.5.0=py37h658377b_0
- cytoolz=0.10.1=py37h516909a_0
- dask=2.17.2=py_0
- dask-core=2.17.2=py_0
- dask-cudf=0.13.0=py37_0
- distributed=2.17.0=py37hc8dfbb8_0
- dlpack=0.2=he1b5a44_1
- double-conversion=3.1.5=he1b5a44_2
- fastavro=0.23.4=py37h8f50634_0
- fastrlock=0.4=py37h3340039_1001
- freetype=2.10.2=he06d7ca_0
- fsspec=0.6.3=py_0
- gflags=2.2.2=he1b5a44_1002
- glog=0.4.0=h49b9bf7_3
- grpc-cpp=1.23.0=h18db393_0
- heapdict=1.0.1=py_0
- icu=64.2=he1b5a44_1
- jinja2=2.11.2=pyh9f0ad1d_0
- joblib=0.15.1=py_0
- jpeg=9c=h14c3975_1001
- ld_impl_linux-64=2.33.1=h53a641e_7
- libblas=3.8.0=16_openblas
- libcblas=3.8.0=16_openblas
- libcudf=0.13.0=cuda10.0_0
- libcuml=0.13.0=cuda10.0_0
- libcumlprims=0.13.0=cuda10.0_0
- libedit=3.1.20181209=hc058e9b_0
- libevent=2.1.10=h72c5cf5_0
- libffi=3.3=he6710b0_1
- libgcc-ng=9.2.0=h24d8f2e_2
- libhwloc=2.1.0=h3c4fd83_0
- libiconv=1.15=h516909a_1006
- liblapack=3.8.0=16_openblas
- libllvm8=8.0.1=hc9558a2_0
- libnvstrings=0.13.0=cuda10.0_0
- libopenblas=0.3.9=h5ec1e0e_0
- libpng=1.6.37=hed695b0_1
- libprotobuf=3.8.0=h8b12597_0
- librmm=0.13.0=cuda10.0_0
- libstdcxx-ng=9.1.0=hdf63c60_0
- libtiff=4.1.0=hfc65ed5_0
- libxml2=2.9.10=hee79883_0
- llvm-openmp=10.0.0=hc9558a2_0
- llvmlite=0.32.0=py37h5202443_0
- locket=0.2.0=py_2
- lz4-c=1.8.3=he1b5a44_1001
- markupsafe=1.1.1=py37h8f50634_1
- msgpack-python=1.0.0=py37h99015e2_1
- nccl=2.6.4.1=hd6f8bf8_0
- ncurses=6.2=he6710b0_1
- numba
- numpy=1.17.5=py37h95a1406_0
- nvstrings=0.13.0=py37_0
- olefile=0.46=py_0
- openssl=1.1.1g=h516909a_0
- packaging=20.4=pyh9f0ad1d_0
- pandas=0.25.3=py37hb3f55d8_0
- parquet-cpp=1.5.1=2
- partd=1.1.0=py_0
- pillow=5.3.0=py37h00a061d_1000
- pip=20.0.2=py37_3
- psutil=5.7.0=py37h8f50634_1
- pyarrow=0.15.0=py37h8b68381_1
- pyparsing=2.4.7=pyh9f0ad1d_0
- python=3.7.7=hcff3b4d_5
- python-dateutil=2.8.1=py_0
- python_abi=3.7=1_cp37m
- pytz=2020.1=pyh9f0ad1d_0
- pyyaml=5.3.1=py37h8f50634_0
- re2=2020.04.01=he1b5a44_0
- readline=8.0=h7b6447c_0
- rmm=0.13.0=py37_0
- setuptools=46.4.0=py37_0
- six=1.15.0=pyh9f0ad1d_0
- snappy=1.1.8=he1b5a44_1
- sortedcontainers=2.1.0=py_0
- sqlite=3.31.1=h62c20be_1
- tblib=1.6.0=py_0
- thrift-cpp=0.12.0=hf3afdfd_1004
- tk=8.6.8=hbc83047_0
- toolz=0.10.0=py_0
- tornado=6.0.4=py37h8f50634_1
- typing_extensions=3.7.4.2=py_0
- ucx=1.7.0+g9d06c3a=cuda10.0_0
- ucx-py=0.13.0+g9d06c3a=py37_0
- ucx-proc=*=gpu
- uriparser=0.9.3=he1b5a44_1
- wheel=0.34.2=py37_0
- xz=5.2.5=h7b6447c_0
- yaml=0.2.4=h516909a_0
- zict=2.0.0=py_0
- zlib=1.2.11=h7b6447c_3
- zstd=1.4.3=h3b9ef0a_0
prefix: /databricks/conda/envs/databricks-ml-gpu
EOF
time conda env update --prefix /databricks/conda/envs/databricks-ml-gpu --file rapids0.13_cuda10.0_ubuntu16.04.yml -vv
time conda install numba=0.48
echo "1" > $INSTALL_FILE
| 0 |
rapidsai_public_repos/cloud-ml-examples | rapidsai_public_repos/cloud-ml-examples/ray/README.md | # RAPIDS Hyperparameter Optimization with Ray Tune
Tune is a scalable hyperparameter optimization (HPO) framework, built
on top of the Ray framework for distributed applications. It includes
modern, scalable HPO algorithms, such as HyperBand and PBT, and it
supports a wide variety of machine learning models.
Ray can run on the public cloud of your choosing, or on on-premise hardware.
RAPIDS integrates smoothly with Ray Tune, using GPU acceleration to
speed up both model training and data prep by up to 40x over CPU-based
alternatives. For HPO sweeps, this can enable you to try more
parameter options and find more accurate classifiers.
## RAPIDS + Ray Tune sample notebooks
This sample notebook shows how to use Ray Tune to optimize XGBoost and
cuML Random Forest classifiers over a large dataset of airline arrival
times. By design, it is very similar to the RAPIDS examples provided
for other cloud and bring-your-own-cloud HPO offerings. As Tune offers
a variety of HPO algorithms, the sample includes utilities to compare
between them. (Note that the "best" HPO algorithm may be *very*
problem-dependent, so results are not fully generalizable.)
You need both Jupyter and RAPIDS 22.08 or later installed to begin. See
https://rapids.ai/start.html for instructions. We recommend using the stable
release. For Ray, you should also install a few additional
packages.
```
pip install tabulate nb_black
pip install -U ray
pip install ray[tune]
pip install bayesian-optimization scikit-optimize
```
## For more details
See the blog post about RAPIDS on Ray Tune (coming soon!).
* For background on the Ray project: https://ray.io/
* To learn more about Ray Tune specifically: https://docs.ray.io/en/latest/tune.html
* cuML documentation for machine learning: https://docs.rapids.ai/api/cuml/nightly/
| 0 |
rapidsai_public_repos/cloud-ml-examples/ray | rapidsai_public_repos/cloud-ml-examples/ray/notebooks/Ray_RAPIDS_HPO.ipynb | # # Uncomment this line to install the packages
# !pip install tabulate nb_black
# !pip install -U ray
# !pip install ray[tune]
# !pip install bayesian-optimization scikit-optimize%load_ext lab_blackimport glob
import logging
import math
import multiprocessing
import os
import subprocess
import sys
import time
from datetime import datetime
from urllib.request import urlretrieve
import cudf
import cuml
import cupy
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import ray
from cuml.model_selection import train_test_split
from sklearn.model_selection import train_test_split as sktrain_test_split
import sklearn
from ray import tune
from ray.tune.experiment import trial
from ray.tune.logger import TBXLogger
from ray.tune.schedulers import AsyncHyperBandScheduler, MedianStoppingRule
_DEBUG = bool(os.environ.get("_DEBUG", False))base_dir = os.getcwd()
num_rows = 2500000 # number of rows to be used in this notebook; max: 115000000
# Ensure the dataset is setup
dataset_config = {
"dataset_name": "airline",
"nrows": num_rows, # Max Rows in dataset: 115000000
"delayed_threshold": 10,
"remote_URL": "http://kt.ijs.si/elena_ikonomovska/datasets/airline/airline_14col.data.bz2",
"local_cache_dir": os.path.join(base_dir, "data"),
}
download_filename = os.path.join(
dataset_config["local_cache_dir"], os.path.basename(dataset_config["remote_URL"])
)
decompressed_filename = os.path.splitext(download_filename)
orc_name = os.path.join(
dataset_config["local_cache_dir"], "airline" + str(dataset_config["nrows"]) + ".orc"
)
data_dir = dataset_config["local_cache_dir"]def prepare_dataset():
global download_filename, decompressed_filename, orc_name
if os.path.isfile(orc_name):
print(f" > File already exists. Ready to load at {orc_name}")
dataset = None
if compute == "GPU":
dataset = cudf.read_orc(orc_name)
elif compute == "CPU":
import pandas as pd
import pyarrow.orc as orc
with open(orc_name, mode="rb") as file:
data = orc.ORCFile(file)
dataset = data.read().to_pandas()
return dataset
# Ensure folder exists
os.makedirs(dataset_config["local_cache_dir"], exist_ok=True)
# download progress tracker
def data_progress_hook(block_number, read_size, total_filesize):
if (block_number % 1000) == 0:
print(
f" > percent complete: { 100 * ( block_number * read_size ) / total_filesize:.2f}\r",
end="",
)
return
if not os.path.exists(download_filename):
print(f"File does not exist, downloading now...")
urlretrieve(
url=dataset_config["remote_URL"],
filename=download_filename,
reporthook=data_progress_hook,
)
print(f" > Download complete {download_filename}")
else:
print(f"Dataset already downloaded")
# Decompressing
completed = subprocess.run(["bzip2", "-d", download_filename])
print("returncode:", completed.returncode)
input_cols = [
"Year",
"Month",
"DayofMonth",
"DayofWeek",
"CRSDepTime",
"CRSArrTime",
"UniqueCarrier",
"FlightNum",
"ActualElapsedTime",
"Origin",
"Dest",
"Distance",
"Diverted",
"ArrDelay",
]
# ensure we respect bounds on rows [ airline max = 115 M ]
nrows = np.min((dataset_config["nrows"], 115000000))
df = cudf.read_csv(decompressed_filename[0], names=input_cols, nrows=nrows)
# turn into binary classification [i.e. flight delays beyond delayed_threshold minutes are considered late ]
df["ArrDelayBinary"] = 1.0 * (df["ArrDelay"] > dataset_config["delayed_threshold"])
# drop non-binary label column [ delay time ]
df = df[df.columns.difference(["ArrDelay"])]
# encode categoricals as numeric
for col in df.select_dtypes(["object"]).columns:
df[col] = df[col].astype("category").cat.codes.astype(np.int32)
# cast all columns to int32
for col in df.columns:
df[col] = df[col].astype(np.float32) # needed for random forest
# put target/label column first [ classic XGBoost standard ]
output_cols = ["ArrDelayBinary"] + input_cols[:-1]
df = df.reindex(columns=output_cols)
df.to_orc(orc_name)
return dftrial_num = 0
def get_trial_name(trial: ray.tune.experiment.trial.Trial):
# Returns the trial number over an iterator variable trail_num
global trial_num
trial_num = trial_num + 1
trial_name = trial.trainable_name + "_" + str(trial_num)
return trial_nameclass PerfTimer:
# High resolution timer for reporting training and inference time.
def __init__(self):
self.start = None
self.duration = None
def __enter__(self):
self.start = time.perf_counter()
return self
def __exit__(self, *args):
self.duration = time.perf_counter() - self.startclass BaseTrainTransformer(tune.Trainable):
@property
def static_config(self) -> dict:
return getattr(self, "_static_config", {})
def setup(self, config: dict):
CSP_paths = {"train_data": data_dir}
if self.static_config["compute"] == "GPU":
self._gpu_id = ray.get_gpu_ids()[0]
self._dataset, self._col_labels, self._y_label = (
ray.get(data_id),
None,
"ArrDelayBinary",
)
# classification objective requires int32 label for cuml random forest
self._dataset[self._y_label] = self._dataset[self._y_label].astype("int32")
self.rf_model = None
self._build(config)
def _build(self, new_config):
self._model_params = {
"max_depth": int(new_config["max_depth"]),
"n_estimators": int(new_config["n_estimators"]),
"max_features": float(new_config["max_features"]),
"n_bins": 16, # args.n_bins,
"seed": time.time(),
}
self._global_best_model = None
self._global_best_test_accuracy = 0
def step(self):
iteration = getattr(self, "iteration", 0)
if self.static_config["compute"] == "GPU":
# split data
X_train, X_test, y_train, y_test = train_test_split(
X=self._dataset,
y=self._y_label,
train_size=0.8,
shuffle=True,
random_state=iteration,
)
self.rf_model = cuml.ensemble.RandomForestClassifier(
n_estimators=self._model_params["n_estimators"],
max_depth=self._model_params["max_depth"],
n_bins=self._model_params["n_bins"],
max_features=self._model_params["max_features"],
)
elif self.static_config["compute"] == "CPU":
# Optionally allow CPU version for performance comparison
X_train, X_test, y_train, y_test = sktrain_test_split(
self._dataset.loc[:, self._dataset.columns != self._y_label],
self._dataset[self._y_label],
train_size=0.8,
shuffle=True,
random_state=iteration,
)
self.rf_model = sklearn.ensemble.RandomForestClassifier(
n_estimators=self._model_params["n_estimators"],
max_depth=self._model_params["max_depth"],
max_features=self._model_params["max_features"],
n_jobs=-1,
)
else:
print("Unknown option. Please select CPU or GPU")
return
# train model
with PerfTimer() as train_timer:
trained_model = self.rf_model.fit(X_train, y_train)
training_time = train_timer.duration
# evaluate perf
with PerfTimer() as inference_timer:
test_accuracy = trained_model.score(X_test, y_test.astype("int32"))
infer_time = inference_timer.duration
# update best model [ assumes maximization of perf metric ]
if test_accuracy > self._global_best_test_accuracy:
self._global_best_test_accuracy = test_accuracy
self._global_best_model = trained_model
return {
"test_accuracy": test_accuracy,
"train_time": round(training_time, 4),
"infer_time": round(infer_time, 4),
"is_bad": not math.isfinite(test_accuracy),
"should_checkpoint": False,
}
def _save(self, checkpoint):
self._global_best_test_accuracy = checkpoint["test_accuracy"]
def _restore(self, checkpoint):
return {
"test_accuracy": self._global_best_test_accuracy,
}
def save_checkpoint(self, checkpoint_dir):
pass
def reset_config(self, new_config):
# Rebuild the config dependent stuff
del self.rf_model
self._build(new_config)
self.config = new_config
return Trueclass WrappedTrainable(BaseTrainTransformer):
def __init__(self, *args, **kwargs):
self._static_config = static_config
super().__init__(*args, **kwargs)def build_search_alg(search_alg, param_ranges: dict):
"""
Initialize a search algorithm that is selected using 'search_alg'
Parameters
----------
search_alg : str; Selecting the search algorithm. Possible values
[BayesOpt, SkOpt]
param_ranges : dictionary of parameter ranges over which the search
should be performed
Returns
-------
alg : Object of the RayTune search algorithm selected
"""
alg = None
if search_alg == "BayesOpt":
from ray.tune.search.bayesopt import BayesOptSearch
alg = BayesOptSearch(
param_ranges,
metric="test_accuracy",
mode="max",
utility_kwargs={"kind": "ucb", "kappa": 2.5, "xi": 0.0},
)
elif search_alg == "SkOpt":
from skopt import Optimizer
from skopt.space import Real, Integer
from ray.tune.search.skopt import SkOptSearch
opt_params = [
Integer(param_ranges["n_estimators"][0], param_ranges["n_estimators"][1]),
Integer(param_ranges["max_depth"][0], param_ranges["max_depth"][1]),
Real(
param_ranges["max_features"][0],
param_ranges["max_features"][1],
prior="log-uniform",
),
]
optimizer = Optimizer(opt_params)
alg = SkOptSearch(
optimizer,
list(param_ranges.keys()),
metric="test_accuracy",
mode="max",
)
else:
print("Unknown Option. Select BayesOpt or SkOpt")
return algdef select_sched_alg(sched_alg):
"""
Initialize a scheduling algorithm that is selected using 'sched_alg'
Parameters
----------
sched_alg : str; Selecting the search algorithm. Possible values
[MedianStop, AsyncHyperBand]
Returns
-------
alg : Object of the RayTune scheduling algorithm selected
"""
sched = None
if sched_alg == "AsyncHyperBand":
sched = AsyncHyperBandScheduler(
time_attr="training_iteration",
metric="test_accuracy",
mode="max",
max_t=50,
grace_period=1,
reduction_factor=3,
brackets=3,
)
elif sched_alg == "MedianStop":
sched = MedianStoppingRule(
time_attr="time_total_s",
metric="test_accuracy",
mode="max",
grace_period=1,
min_samples_required=3,
)
else:
print("Unknown Option. Select MedianStop or AsyncHyperBand")
return schednum_samples = 50
compute = (
"GPU" # Can take a CPU value (only for performance comparison. Not recommended)
)
CV_folds = 3 # The number of Cross-Validation folds to be performed
search_alg = "SkOpt" # Options: SkOpt or BayesOpt
sched_alg = "AsyncHyperBand" # Options: AsyncHyperBand or MedianStop
max_concurrent = 10 # Number of concurrent samples; if -1, determined by number of available GPU for GPU tasks.
# HPO Param ranges
# NOTE: Depending on the GPU memory we might need to adjust the parameter range for a successful run
n_estimators_range = (50, 1000)
max_depth_range = (2, 15)
max_features_range = (0.1, 0.8)
hpo_ranges = {
"n_estimators": n_estimators_range,
"max_depth": max_depth_range,
"max_features": max_features_range,
}if _DEBUG:
# Only use 1 GPU when debugging
ray.init(local_mode=True, num_gpus=1)
num_samples = 3
else:
ray.init(dashboard_host="0.0.0.0")
if max_concurrent == -1:
if compute == "GPU":
max_concurrent = cupy.cuda.runtime.getDeviceCount()
else:
raise Exception("For CPU, must specify max_concurrent value")
cpu_per_sample = int(multiprocessing.cpu_count() / max_concurrent)cdf = prepare_dataset()
# for shared access across processes
data_id = ray.put(cdf)search = build_search_alg(search_alg, hpo_ranges)
sched = select_sched_alg(sched_alg)exp_name = None
if exp_name is not None:
exp_name += exp_name
else:
exp_name = ""
exp_name += "{}_{}_CV-{}_{}M_SAMP-{}".format(
"RF", compute, CV_folds, int(num_rows / 1000000), num_samples
)
exp_name += "_{}".format("Random" if search_alg is None else search_alg)
if sched_alg is not None:
exp_name += "_{}".format(sched_alg)
static_config = {
"dataset_filename": os.path.basename(orc_name),
"compute": compute,
"num_workers": cpu_per_sample if compute == "CPU" else 1,
}%tb
results_dir = "results"
analysis = tune.run(
WrappedTrainable,
name=exp_name,
scheduler=sched,
search_alg=search,
stop={
"training_iteration": CV_folds,
"is_bad": True,
},
resources_per_trial={"cpu": cpu_per_sample, "gpu": int(compute == "GPU")},
num_samples=num_samples,
checkpoint_at_end=True,
keep_checkpoints_num=1,
local_dir=results_dir,
trial_name_creator=get_trial_name,
checkpoint_score_attr="test_accuracy",
verbose=1,
raise_on_failed_trial=False,
)# Save results for plotting
csv_path = os.path.join(os.getcwd(), results_dir, f"{search_alg}_trials.csv")
analysis.dataframe().to_csv(csv_path)from scipy.interpolate.interpolate import interp1d
from scipy.signal import savgol_filter
import matplotlib.lines as mlines
import matplotlib.patches as mpatches
import matplotlib.markers as mmarkers
trials_glob = os.path.join(os.getcwd(), results_dir, "*.csv")
trials_csvs = glob.glob(trials_glob)
csvs_by_type = {}
for trial in trials_csvs:
csv_name = os.path.split(trial)[-1]
trial_type = csv_name.split("_")[0]
if trial_type not in csvs_by_type:
csvs_by_type[trial_type] = []
csvs_by_type[trial_type].append(trial)combined_dfs = {}
max_val = 0.0
min_val = 1.0
n_rows = num_samples
for type_name, trials in csvs_by_type.items():
temp_dfs = []
max_val = 0.0
min_val = 1.0
for trial in trials:
imported_trial = pd.read_csv(trial)
filtered_df: pd.DataFrame = imported_trial["test_accuracy"]
max_val = max(max_val, filtered_df.values.max())
min_val = min(min_val, filtered_df.values.min())
temp_dfs.append(filtered_df)
comb_df = pd.concat(temp_dfs, axis=1)
mean = comb_df.mean(axis=1)
std = comb_df.std(axis=1)
max_col = comb_df.max(axis=1)
min_col = comb_df.min(axis=1)
upper = pd.concat([mean + std, max_col], axis=1).min(axis=1)
lower = pd.concat([mean - std, min_col], axis=1).max(axis=1)
cummax = comb_df.cummax()
comb_df["Mean"] = mean
comb_df["Std"] = std
comb_df["Max"] = max_col
comb_df["Min"] = min_col
comb_df["Upper"] = upper
comb_df["Lower"] = lower
# comb_df["CumMax"] = cummax
combined_dfs[type_name] = comb_df
if len(comb_df) < n_rows:
n_rows = len(comb_df)
print(
"{} (min, max): {}, {}".format(type_name, round(min_val, 5), round(max_val, 5))
)import seaborn as sns
x = np.arange(0, n_rows, 1)
x_smoothed = np.linspace(x.min(), x.max(), 150)
sns.set_style("darkgrid")
fig, axes = None, None
if len(combined_dfs) == 2:
fig, axes = plt.subplots(4, 1, sharex=True, sharey=True, figsize=(10, 28))
elif len(combined_dfs) == 1:
fig, axes = plt.subplots(3, 1, sharex=True, sharey=True, figsize=(10, 28))
all_axes = {}
i = 0
for key in combined_dfs.keys():
all_axes[key] = axes[i]
i += 1
all_axes["All"] = axes[i]
i += 1
all_axes["Max"] = axes[i]
comb_axes = axes[i - 1]
max_axes = axes[i]
colors = {
"BayesOpt": "C2",
"SkOpt": "C1",
}
comb_legend_handles = []
max_legend_handles = []
for name, df in combined_dfs.items():
df = df[:n_rows]
ymean = df["Mean"]
ystd = df["Std"]
ystd_up = df["Upper"]
ystd_dn = df["Lower"]
ax = all_axes[name]
# Smooth the means to make it easier to read
itp = interp1d(x, ymean, kind="linear")
window_size, poly_order = 17, 3
ymean_smoothed = savgol_filter(itp(x_smoothed), window_size, poly_order)
ax.plot(x_smoothed, ymean_smoothed, "-", color=colors[name])
ax.plot(x, df["test_accuracy"], "o", color=colors[name], markersize=1.0)
ax.fill_between(x, ystd_up, ystd_dn, alpha=0.2, color=colors[name])
ax.set_title(name)
mean_line = mlines.Line2D(
[], [], color=colors[name], markersize=1, label="Smoothed Mean"
)
points = mlines.Line2D(
[],
[],
color=colors[name],
marker="o",
markersize=1,
linestyle="",
label="Achieved Accuracy",
)
std_patch = mpatches.Patch(color=colors[name], alpha=0.2, label="Std. Dev")
ax.legend(handles=[mean_line, points, std_patch], loc="lower right")
comb_axes.set_title("Comparison of Average Output Per Optimization")
comb_axes.plot(x_smoothed, ymean_smoothed, "-", color=colors[name])
comb_axes.plot(x, ymean, "o", color=colors[name], markersize=2)
comb_axes.fill_between(x, ystd_up, ystd_dn, alpha=0.2, color=colors[name])
comb_legend_handles.append(
mlines.Line2D([], [], color=colors[name], marker="o", markersize=2, label=name)
)
max_axes.set_title("Comparison of Average Cumulative Max Value Per Optimization")
max_axes.plot(
x,
df[["test_accuracy"]].cummax().mean(axis=1),
"-",
color=colors[name],
alpha=1.0,
)
max_legend_handles.append(mlines.Line2D([], [], color=colors[name], label=name))
comb_axes.legend(handles=comb_legend_handles, loc="lower right")
max_axes.legend(handles=max_legend_handles, loc="lower right")
fig.tight_layout()
fig.savefig("SearchComparison.png", facecolor="white", transparent=False) | 0 |
rapidsai_public_repos/cloud-ml-examples | rapidsai_public_repos/cloud-ml-examples/gcp/README.md | # **This guide is deprecated an no longer maintained.**
## Quick start guide
Here we will go over some common tasks, related to utilizing RAPIDS on the GCP AI Platform. Note that strings containing '[YOUR_XXX]' indicate items that you will need to supply, based on your specific resource names and environment.
### Create a Notebook using the RAPIDS environemnt
Motivation: We would like to create a GCP notebook with RAPIDS 0.18 release
Workflow: We will create a notebook instance using the `RAPIDS 0.18 [Experimental]` env
1. Log into your GCP console.
1. Select AI-Platform -> Notebooks
1. Select a "New Instance" -> "RAPIDS 0.18 [Experimental]"
1. Select 'Install NVIDIA GPU driver automatically for me'
1. Create
1. Once JupterLab is running, you will have jupyter notebooks with rapids installed and rapids notebook examples under tutorials/RapidsAi.
To create an instance with A100s:
1. Select "New Instance" -> "Customize instance"
1. Select us-central1 region
1. Select "RAPIDS 0.18 [Experimental]" Environment
1. Choose A2 highgpu (for 1, 2 4 and 8 A100s) or A1 megagpu (16x A100s) as machine type
### Install RAPIDS on a pre-made Notebook
Motivation: We have an existing GCP notebook that we wish to update to support RAPIDS functionality.
Workflow: We will create a notebook instance, and run a shell script that will install a Jupyter kernel and allow us to run RAPIDS based tasks.
1. Log into your GCP console.
1. Select AI-Platform -> Notebooks
1. Select a "New Instance" -> "Python 3 (CUDA Toolkit 11.0)" -> With 1 NVIDIA Tesla T4
1. Select 'Install NVIDIA GPU driver automatically for me'
1. Create.
1. Once JupyterLab is running
1. Open a new terminal
1. Run
```shell
RAPIDS_VER=21.06
CUDA_VER=11.0
wget -q https://data.rapids.ai/conda-pack/rapidsai/rapids${RAPIDS_VER}_cuda${CUDA_VER}_py3.8.tar.gz
tar -xzf rapids${RAPIDS_VER}_cuda${CUDA_VER}_py3.8.tar.gz -C /opt/conda/envs/rapids_py38
conda activate rapids_py38
conda unpack
ipython kernel install --user --name=rapids_py38
```
1. Once completed, you will now have a new kernel in your jupyter notebooks called 'rapids_py38' which will have rapids installed.
### Deploy a custom RAPIDS training container utilizing the 'airline dataset', and initiate a training job with support for HyperParameter Optimization (HPO)
Motivation: We would like to be able to utilize GCP's AI Platform for training a custom model, utilizing RAPIDS.
Workflow: Install the required libraries, and authentication components for GCP, configure a storage bucket for persistent data, build our custom training container, upload the container, and launch a training job with HPO.
1. Install GCP 'gcloud' SDK
1. See: https://cloud.google.com/sdk/install
1. Configure gcloud authorization for docker on your build machine
1. See: https://cloud.google.com/container-registry/docs/advanced-authentication
1. Configure a google cloud object storage bucket that will provide and output location
1. Pull or build training containers and upload to GCR
1. Pull
1. Find the appropriate container: [Here](https://hub.docker.com/r/rapidsai/rapidsai-cloud-ml/tags?page=0&ordering=last_updated)
1. `docker tag <image> gcr.io/[YOUR_PROJECT_NAME]/rapids_training_container:latest`
1. Build
1. `$ cd .`
1. `$ docker build --tag gcr.io/[YOUR_PROJECT_NAME]/rapids_training_container:latest --file common/docker/Dockerfile.training.unified .`
1. `$ docker push gcr.io/[YOUR_PROJECT_NAME]/rapids_training_container:latest`
1. Training via GCP UI
1. A quick note regarding GCP's cloudml-Hypertune
1. This library interacts with the GCP AI Platform's HPO process by reporting required optimization metrics to the system after each training iteration.
```python
hpt.report_hyperparameter_tuning_metric(
hyperparameter_metric_tag='hpo_accuracy',
metric_value=accuracy)
```
1. For our purposes, the 'hyperparameter_metric_tag' should always correspond to the 'Metric to optimize' element passed to a job deployment.
1. Training Algorithm
1. From the GCP console select 'jobs' -> 'new training job' -> custom code training
1. Choose 'Select a container image from the container Registry'
1. Set 'Master image' to 'gcr.io/[YOUR_PROJECT_NAME]/rapids_training_container:latest'
1. Set 'Job directory' to 'gs://[YOUR_GOOGLE_STORAGE_BUCKET]'
1. Algorithm Arguments
1. Ex:
1. ```bash
--train
--do-hpo
--cloud-type=GCP
--data-input-path=gs://[YOUR STORAGE BUCKET]
--data-output-path=gs://[YOUR STORAGE BUCKET]/training_output
--data-name=airline_20000000.orc
```
1. With Hypertune
1. Enter the hypertune parameters. Ex:
1. ```bash
Argument name: hpo-max-depth
Type: Integer
Min: 2
Max: 8
1. ```bash
Argumnet name: hpo-num-est
Type: Integer
Min: 100
Max: 200
```
1. ```bash
Argument name: hpo-max-features
Type: Double
Min: 0.2
Max: 0.6
```
1. Enter an optimizing metric. Ex:
1. ```bash
Metric to optimize: hpo_accuracy
Goal: Maximize
Max trials: 20
Max parallel trials: 5
Algorithm: Bayesian optimization
Early stopping: True
```
1. Job Settings
1. ```bash
Job ID: my-test-job
Region: us-central1
1. Scale Tier
1. Select 'CUSTOM' -> 'Use Compute Engine Machine Types'
1. Master Node
1. Ex. n1-standard-8
1. Accelerator
1. Ex. V100 or T4. K80s are not supported.
1. Select 'Done', and launch your training job.
1. Training via gcloud job submission
1. Update your training configuration based on 'example_config.json'
1. ```json
{
"trainingInput": {
"args": [
"--train",
"--do-hpo",
"--cloud-type=GCP",
"--data-input-path=gs://[YOUR STORAGE BUCKET]",
"--data-output-path=gs://[YOUR STORAGE BUCKET]/training_output",
"--data-name=airline_20000000.orc"
],
"hyperparameters": {
"enableTrialEarlyStopping": true,
"goal": "MAXIMIZE",
"hyperparameterMetricTag": "hpo_accuracy",
"maxParallelTrials": 1,
"maxTrials": 2,
"params": [
{
"maxValue": 200,
"minValue": 100,
"parameterName": "hpo-num-est",
"type": "INTEGER"
},
{
"maxValue": 17,
"minValue": 9,
"parameterName": "hpo-max-depth",
"type": "INTEGER"
},
{
"maxValue": 0.6,
"minValue": 0.2,
"parameterName": "hpo-max-features",
"type": "DOUBLE"
}
]
},
"jobDir": "gs://[YOUR PROJECT NAME]/training_output",
"masterConfig": {
"imageUri": "gcr.io/[YOUR PROJECT NAME]/rapids_training_container:latest",
"acceleratorConfig": {
"count": "1",
"type": "NVIDIA_TESLA_T4"
}
},
"masterType": "n1-standard-8",
"region": "us-west1",
"scaleTier": "CUSTOM"
}
}
1. For more information, see:
1. https://cloud.google.com/sdk/gcloud/reference/ai-platform/jobs/submit/training
1. Run your training job
1. `$ gcloud ai-platform jobs submit training [YOUR_JOB_NAME] --config ./example_config.json`
1. Monitor your training job
1. `$ gcloud ai-platform jobs stream-logs [YOUR_JOB_NAME]`
| 0 |
rapidsai_public_repos/cloud-ml-examples/gcp | rapidsai_public_repos/cloud-ml-examples/gcp/notebook_setup/README.md | # **This guide is deprecated an no longer maintained.**
## **Pack and Deploy Conda Environments for RAPIDS on Google Cloud Platform (GCP)**
This section describes the process required to:
1. Package and deploy a RAPIDS conda environment via helper script
1. Package and deploy a RAPIDS conda environment manually
1. Initialize a RAPIDS conda environment.
1. Package the environment using conda-pack.
1. Unpack into a second second environment with conda-unpack.
1. Unpack an existing conda environment via cloud storage.
### **Package and Deploy Using the Helper Script**
#### Pack Environment
1. `common/code/create_packed_conda_env`
```bash
... processing ...
Packing conda environment
Collecting packages...
Packing environment at '[CONDA ENV]/rapids21.06_py3.8' to 'rapids21.06_py3.8.tar.gz'
[########################################] | 100% Completed | 1min 51.1s
```
#### Unpack Environment on Target System
1. Copy your environment tarball (rapids_py38.tar.gz) to the target system
1. Unpack in desired environment
1. `common/code/create_packed_conda_env --action unpack`
1. Alternatively, the environment can be manually unpacked as
```bash
CONDA_ENV="rapids21.06_py3.8"
TARBALL="$CONDA_ENV.tar.gz"
UNPACK_TO="$CONDA_ENV"
mkdir -p "$UNPACK_TO"
tar -xzf $TARBALL -C "$UNPACK_TO"
source "$UNPACK_TO/bin/activate"
conda-unpack
python -m ipykernel install --user --name $CONDA_ENV
```
### **Package and Deploy Manually**
#### Pack Environment
1. Create a clean conda environment to work from
1. `conda create --name=rapids_env python=3.7`
1. Install conda-pack
1. `conda install -c conda-forge conda-pack`
1. Install RAPIDS
1. Select the package level to install from [RAPIDS.ai](rapids.ai/start.html)
1. Ex. For a full install on Ubuntu 18.04, with CUDA 11.2
```bash
conda install -c rapidsai -c nvidia -c conda-forge
rapids=21.06 python=3.8 cudatoolkit=11.2
```
1. Pack your environment
1. `conda-pack -n rapids_env -o rapids_py38.tar.gz`
#### Unpack Environment on Target System
1. Copy your environment tarball (rapids_py38.tar.gz) to the target system
1. Extract the tarball to the desired environment
1. Ex. Local
```bash
mkdir -p ./rapids_env
tar -xzf rapids_py38.tar.gz -C ./rapids_env
source ./rapids_env/bin/activate
```
1. Ex. Anaconda
```bash
mkdir -p $HOME/anaconda3/envs/rapids_py38
tar -xzf rapids_py38.tar.gz -C $HOME/anaconda3/envs/rapids_py38
source $HOME/anaconda3/envs/rapids_py38/bin/activate
```
1. Cleanup environment prefixes
1. `conda-unpack`
### **Unpack an Existing Environment Via Cloud Storage**
#### Unpacking on a Target Environment
1. Upload your packed conda environment to a GCP storage bucket.
1. Pull and unpack your environment manually or via script as
```bash
GCP_STORAGE="https://storage.googleapis.com/$YOUR_BUCKET/rapids_py38.tar.gz"
CONDA_ENV="rapids21.06_py3.8"
TARBALL="$CONDA_ENV.tar.gz"
UNPACK_TO="$CONDA_ENV"
wget -q --show-progress $GCP_STORAGE
mkdir -p "$UNPACK_TO"
tar -xzf $TARBALL -C "$UNPACK_TO"
source "$UNPACK_TO/bin/activate"
conda-unpack
python -m ipykernel install --user --name $CONDA_ENV
```
| 0 |
rapidsai_public_repos/cloud-ml-examples/gcp | rapidsai_public_repos/cloud-ml-examples/gcp/notebooks/container_build.ipynb | ## GCLOUD_BIN_PATH=[path to the location where 'gcloud' bin is installed]
## See: https://cloud.google.com/sdk/install
import json
import os
import subprocess
GCLOUD_BIN_PATH = "[/path/to/gcloud/location]"
GCP_PROJECT_NAME = "[YOUR PROJECT NAME]"
GCP_STORAGE_PATH = "[PATH TO GCP STORAGE LOCATION]" # Ex. gs://[path_to_your_data]/subdir
gcloud_env = os.environ.copy()
gcloud_env["PATH"] = f"{gcloud_env['PATH']}:{GCLOUD_BIN_PATH}"
def exec_cmd_and_return(*popenargs, **kwargs):
process = subprocess.Popen(stdout=subprocess.PIPE, stderr=subprocess.PIPE,
env=gcloud_env, *popenargs, **kwargs)
output, err = process.communicate()
return_code = process.poll()
output = str(output.decode())
return (return_code, output + str(err))
command = "docker build -t rapids_training_test --file Dockerfile.training ./".split()
_, result = exec_cmd_and_return(command)
print(result)
command = f"docker tag rapids_training_test:latest gcr.io/{GCP_PROJECT_NAME}/rapids_training_container:latest".split()
_, result = exec_cmd_and_return(command)
print(result)command = f"docker push gcr.io/{GCP_PROJECT_NAME}/rapids_training_container:latest".split()
_, result = exec_cmd_and_return(command)
print(result)
command = "gcloud auth configure-docker".split()
_, result = exec_cmd_and_return(command)
print(result)config_name = "gcloud_training_config.json"
config = {
"trainingInput": {
"args": [
"--train",
"--do-hpo",
"--hpo-num-bins=64",
"--cloud-type=GCP",
"--compute-type=GPU",
f"--data-input-path=gs://{GCP_STORAGE_PATH}",
f"--data-output-path=gs://{GCP_STORAGE_PATH}/training_output",
"--data-name=airline_20000000.orc",
"--model-type=RandomForest"
],
"hyperparameters": {
"enableTrialEarlyStopping": True,
"goal": "MAXIMIZE",
"hyperparameterMetricTag": "hpo_accuracy",
"maxParallelTrials": 1,
"maxTrials": 1,
"maxFailedTrials": 1,
"params": [
{
"maxValue": 600,
"minValue": 100,
"parameterName": "hpo-num-est",
"type": "INTEGER"
},
{
"maxValue": 20,
"minValue": 9,
"parameterName": "hpo-max-depth",
"type": "INTEGER"
},
{
"maxValue": 0.6,
"minValue": 0.2,
"parameterName": "hpo-max-features",
"type": "DOUBLE"
}
]
},
"jobDir": f"gs://{GCP_STORAGE_PATH}/training_output",
"masterConfig": {
"imageUri": f"gcr.io/{GCP_PROJECT_NAME}/rapids_training_container:latest",
"acceleratorConfig": {
"count": "1",
"type": "NVIDIA_TESLA_T4"
}
},
"masterType": "n1-standard-8",
"region": "us-west1",
"scaleTier": "CUSTOM"
}
}
with open(config_name, 'w') as writer:
writer.write(json.dumps(config, indent=4, sort_keys=True))
experiment_name = "test_experiment_01"
command = f"gcloud ai-platform jobs submit training {experiment_name} --config ./{config_name}".split()
_, result = exec_cmd_and_return(command)
print(_)
print(result)
| 0 |
rapidsai_public_repos/cloud-ml-examples/gcp | rapidsai_public_repos/cloud-ml-examples/gcp/notebooks/custom_hpo.ipynb | ### Configure environmentimport json
import logging
import random
import sys
from ax import ParameterType, optimize# os
import sys, os, time, logging
# CPU DS stack
import pandas as pd
import numpy as np
import sklearn
# GPU DS stack [ rapids ]
import gcsfs
# scaling library
import dask
# data ingestion [ CPU ]
from pyarrow import orc as pyarrow_orc
# ML models
from sklearn import ensemble
import xgboost
# data set splits
from sklearn.model_selection import train_test_split as sklearn_train_test_split
# device query
##hack
try:
import cudf, cuml
from cuml.model_selection import train_test_split as cuml_train_test_split
import pynvml
import cupy
except Exception as e:
print("Caught import failures -- probably missing GPU:")
print(e)
# memory query
import psutil
# i/o
import logging, json, pprint
default_sagemaker_paths = {
'base': '/opt/ml',
'code': '/opt/ml/code',
'data': '/opt/ml/input',
'train_data': '/opt/ml/input/data/training',
'hyperparams': '/opt/ml/input/config/hyperparameters.json',
'model': '/opt/ml/model',
'output': '/opt/ml/output',
}
class RapidsCloudML(object):
def __init__(self, cloud_type='AWS',
model_type='XGBoost',
data_type='ORC',
compute_type='single-GPU',
n_workers=-1,
verbose_estimator=False,
CSP_paths=default_sagemaker_paths):
self.CSP_paths = CSP_paths
self.cloud_type = cloud_type
self.model_type = model_type
self.data_type = data_type
self.compute_type = compute_type
self.verbose_estimator = verbose_estimator
self.n_workers = self.parse_compute(n_workers)
self.query_memory()
def _read_orc(self, filename):
if ('CPU' in self.compute_type):
if (filename.startswith('gs://')):
fs = gcsfs.GCSFileSystem()
with fs.open(filename, mode='rb') as file:
dataset = pyarrow_orc.ORCFile(file).read().to_pandas()
else:
with open(filename, mode='rb') as file:
dataset = pyarrow_orc.ORCFile(file).read().to_pandas()
elif ('GPU' in self.compute_type):
dataset = cudf.read_orc(filename)
return dataset
def _read_csv(self, filename, col_labels):
if ('CPU' in self.compute_type):
dataset = pd.read_csv(filename, names=col_labels)
elif ('GPU' in self.compute_type):
dataset = cudf.read_csv(filename, names=col_labels)
return dataset
def load_data(self, filename='dataset.orc', col_labels=None, y_label='ArrDelayBinary'):
target_filename = self.CSP_paths['train_data'] + '/' + filename
self.log_to_file(f'\n> loading dataset from {target_filename}...\n')
with PerfTimer() as ingestion_timer:
if 'ORC' in self.data_type:
dataset = self._read_orc(target_filename)
elif 'CSV' in self.data_type:
dataset = self._read_csv(target_filename, names=col_labels)
self.log_to_file(f'ingestion completed in {ingestion_timer.duration}')
self.log_to_file(f'dataset descriptors: {dataset.shape}\n {dataset.dtypes}\n {dataset.columns}\n')
return dataset, col_labels, y_label, ingestion_timer.duration
def split_data(self, dataset, y_label, train_size=.8, random_state=0, shuffle=True):
"""
split dataset into train and test subset
NOTE: assumes the first column of the dataset is the classification labels
! in the case of sklearn, we manually filter this column in the split call
! in the case of cuml, the filtering happens internally
"""
self.log_to_file('\tsplitting train and test data')
start_time = time.perf_counter()
with PerfTimer() as split_timer:
if 'CPU' in self.compute_type:
X_train, X_test, y_train, y_test = sklearn_train_test_split(dataset.loc[:, dataset.columns != y_label],
dataset[y_label], train_size=train_size,
shuffle=shuffle, random_state=random_state)
elif 'GPU' in self.compute_type:
X_train, X_test, y_train, y_test = cuml_train_test_split(X=dataset, y=y_label, train_size=train_size,
shuffle=shuffle, random_state=random_state)
self.log_to_file(f'\t> split completed in {split_timer.duration}')
return X_train, X_test, y_train, y_test, split_timer.duration
def train_model(self, X_train, y_train, model_params):
self.log_to_file(f'\ttraining {self.model_type} estimator w/ hyper-params')
pprint.pprint(model_params, indent=10)
print(f"model type: {self.model_type}\n compute type: {self.compute_type}\n dataset dtype: {type(X_train)}")
try:
if self.model_type == 'XGBoost':
trained_model, training_time = self.fit_xgboost(X_train, y_train, model_params)
elif self.model_type == 'RandomForest':
trained_model, training_time = self.fit_random_forest(X_train, y_train, model_params)
except Exception as error:
self.log_to_file('!error during model training: ' + str(error))
raise
self.log_to_file(f'\t> finished training in {training_time:.4f} s')
return trained_model, training_time
# train dlmc.xgboost model
def fit_xgboost(self, X_train, y_train, model_params):
with PerfTimer() as train_timer:
train_DMatrix = xgboost.DMatrix(data=X_train, label=y_train)
trained_model = xgboost.train(dtrain=train_DMatrix,
params=model_params,
num_boost_round=model_params['num_boost_round'],
verbose_eval=self.verbose_estimator)
return trained_model, train_timer.duration
# fit_xgboost_multi_GPU ()
# fit_random_forest_multi_GPU ()
# train cuml.random-forest model
def fit_random_forest(self, X_train, y_train, model_params):
if 'CPU' in self.compute_type:
rf_model = sklearn.ensemble.RandomForestClassifier(n_estimators=model_params['n_estimators'],
max_depth=model_params['max_depth'],
max_features=model_params['max_features'],
n_jobs=int(self.n_workers),
verbose=self.verbose_estimator)
elif 'GPU' in self.compute_type:
rf_model = cuml.ensemble.RandomForestClassifier(n_estimators=model_params['n_estimators'],
max_depth=model_params['max_depth'],
n_bins=model_params['n_bins'],
max_features=model_params['max_features'],
verbose=self.verbose_estimator)
with PerfTimer() as train_timer:
trained_model = rf_model.fit(X_train, y_train)
return trained_model, train_timer.duration
def evaluate_test_perf(self, trained_model, X_test, y_test):
self.log_to_file(f'\tinferencing on test set')
with PerfTimer() as inference_timer:
try:
if self.model_type == 'XGBoost':
test_DMatrix = xgboost.DMatrix(data=X_test, label=y_test)
test_accuracy = 1 - float(trained_model.eval(test_DMatrix).split(':')[1])
elif self.model_type == 'RandomForest':
# y_test = cudf.DataFrame({'label': y_test.astype('int32') })
test_accuracy = trained_model.score(X_test, y_test.astype('int32'))
except Exception as error:
self.log_to_file('!error during inference: ' + str(error))
raise
self.log_to_file(f'\t> finished inference in {inference_timer.duration:.4f} s')
return test_accuracy, inference_timer.duration
# TODO: FIL inference [ ? ]
# evaluate_perf_FIL(self, trained_model, X_test, y_test ):
# TODO: global_best_model.save()
def save_best_model(self, global_best_model=None):
pass
# ------------------------------------------------------
# end of data science logic
# ------------------------------------------------------
def parse_compute(self, n_workers=None):
if 'CPU' in self.compute_type or 'GPU' in self.compute_type:
available_devices = self.query_compute()
if n_workers == -1:
n_workers = available_devices
assert (n_workers <= available_devices)
self.log_to_file(f'compute type: {self.compute_type}, n_workers: {n_workers}')
else:
raise Exception('unsupported compute type')
return n_workers
def query_compute(self):
available_devices = None
if 'CPU' in self.compute_type:
available_devices = os.cpu_count()
self.log_to_file(f'detected {available_devices} CPUs')
elif 'GPU' in self.compute_type:
available_devices = cupy.cuda.runtime.getDeviceCount()
self.log_to_file(f'detected {available_devices} GPUs')
return available_devices
# TODO: enumerate all visible GPUs [ ? ]
def query_memory(self):
def print_device_memory(memory, device_ID=-1):
memory_free_GB = np.array(memory.free) / np.array(10e8)
memory_used_GB = np.array(memory.used) / np.array(10e8)
memory_total_GB = np.array(memory.total) / np.array(10e8)
if device_ID != -1:
self.log_to_file(f'device ID = {device_ID}')
self.log_to_file(f'memory free, used, total: {memory_free_GB}, {memory_used_GB}, {memory_total_GB}')
if 'CPU' in self.compute_type:
print_device_memory(psutil.virtual_memory())
elif 'GPU' in self.compute_type:
pynvml.nvmlInit()
for iGPU in range(self.n_workers):
handle = pynvml.nvmlDeviceGetHandleByIndex(iGPU)
print_device_memory(pynvml.nvmlDeviceGetMemoryInfo(handle))
def set_up_logging(self):
logging_path = self.CSP_paths['output'] + '/log.txt'
logging.basicConfig(filename=logging_path,
level=logging.INFO)
def log_to_file(self, text):
logging.info(text)
print(text)
def environment_check(self):
self.check_dirs()
if self.cloud_type == 'AWS':
try:
self.list_files('/opt/ml')
self.log_to_file(os.environ['SM_NUM_GPUS'])
self.log_to_file(os.environ['SM_TRAINING_ENV'])
self.log_to_file(os.environ['SM_CHANNEL_TRAIN'])
self.log_to_file(os.environ['SM_HPS'])
except:
pass
else:
pass
def check_dirs(self):
self.log_to_file('\n> checking for sagemaker paths...\n')
directories_to_check = self.CSP_paths
for iDir, val in directories_to_check.items():
self.log_to_file(f'{val}, exists : {os.path.exists(val)}')
self.log_to_file(f'working directory = {os.getcwd()}')
def list_files(self, startpath):
print(f'\n> listing contents of {startpath}\n')
for root, dirs, files in os.walk(startpath):
level = root.replace(startpath, '').count(os.sep)
indent = ' ' * 4 * (level)
print('{}{}/'.format(indent, os.path.basename(root)))
subindent = ' ' * 4 * (level + 1)
for f in files:
print('{}{}'.format(subindent, f))
# perf_counter = highest available timer resolution
class PerfTimer:
def __init__(self):
self.start = None
self.duration = None
def __enter__(self):
self.start = time.perf_counter()
return self
def __exit__(self, *args):
self.duration = time.perf_counter() - self.start
'''
https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html#sklearn.ensemble.RandomForestClassifier.fit
n_estimators=100,
criterion='gini',
max_depth=None,
min_samples_split=2,
min_samples_leaf=1,
min_weight_fraction_leaf=0.0,
max_features='auto',
max_leaf_nodes=None,
min_impurity_decrease=0.0,
min_impurity_split=None,
bootstrap=True,
oob_score=False,
n_jobs=None,
random_state=None,
verbose=0,
warm_start=False,
class_weight=None,
ccp_alpha=0.0,
max_samples=None
'''
### Setup paths, model, and configuration parameters.def gcp_path_setup(input_path, output_path):
paths = {
'train_data': input_path,
'model': f'{output_path}/model',
'output': f'{output_path}/output',
}
return paths
def ax_train_proxy(model_params, config_params, ax_params):
rcml = RapidsCloudML(cloud_type=config_params['cloud_type'],
model_type=config_params['model_type'],
compute_type=f"single-{config_params['compute']}",
CSP_paths=config_params['paths'])
# environment check
rcml.environment_check()
# ingest data [ post pre-processing ]
dataset, col_labels, y_label, ingest_time = rcml.load_data(filename=config_params['dataset_filename'])
rcml.query_memory()
# classification objective requires int32 label for cuml random forest
dataset[y_label] = dataset[y_label].astype('int32')
accuracy_per_fold = []
train_time_per_fold = []
infer_time_per_fold = []
split_time_per_fold = []
global_best_model = None
global_best_test_accuracy = 0
model_params["max_depth"] = ax_params["max_depth"]
model_params["max_features"] = ax_params["max_features"]
model_params["n_estimators"] = ax_params["n_estimators"]
# optional cross-validation w/ model_params['n_train_folds'] > 1
for i_train_fold in range(config_params['CV_folds']):
print(f"STARTING TRAINING FOLD {i_train_fold}", flush=True)
rcml.log_to_file(f"\n CV fold {i_train_fold} of {config_params['CV_folds']}\n")
# split data
X_train, X_test, y_train, y_test, split_time = rcml.split_data(dataset=dataset,
y_label=y_label,
random_state=i_train_fold,
shuffle=True)
split_time_per_fold += [round(split_time, 4)]
# train model
trained_model, training_time = rcml.train_model(X_train, y_train, model_params)
train_time_per_fold += [round(training_time, 4)]
# evaluate perf
test_accuracy, infer_time = rcml.evaluate_test_perf(trained_model, X_test, y_test)
accuracy_per_fold += [round(test_accuracy, 4)]
infer_time_per_fold += [round(infer_time, 4)]
# update best model [ assumes maximization of perf metric ]
if test_accuracy > global_best_test_accuracy:
global_best_test_accuracy = test_accuracy
global_best_model = trained_model
rcml.log_to_file(f'\n accuracy per fold : {accuracy_per_fold} \n')
rcml.log_to_file(f'\n train-time per fold : {train_time_per_fold} \n')
rcml.log_to_file(f'\n infer-time per fold : {infer_time_per_fold} \n')
rcml.log_to_file(f'\n split-time per fold : {split_time_per_fold} \n')
return global_best_test_accuracy
def ax_train(model_params: dict, config_params: dict):
depth = [int(d) for d in config_params['ht_depth_range'].split(',')]
features = [float(d) for d in config_params['ht_features_range'].split(',')]
estimators = [int(d) for d in config_params['ht_est_range'].split(',')]
experiments = config_params['ht_experiments']
parameters=[
{"name": "max_depth", "type": "range", "bounds": depth, "parameter_type": ParameterType.INT},
{"name": "max_features", "type": "range", "bounds": features, "parameter_type": ParameterType.FLOAT},
{"name": "n_estimators", "type": "range", "bounds": estimators, "parameter_type": ParameterType.INT}
]
best_parameters, best_values, experiment, model = optimize(
parameters=parameters,
evaluation_function=lambda params: ax_train_proxy(model_params=model_params,
config_params=config_params,
ax_params=params),
minimize=False,
total_trials=experiments,
objective_name='accuracy',
)
print("Ax Optimization Results:")
print(best_parameters)
print(best_values)
return best_values['accuracy']paths = gcp_path_setup("gs://[PATH TO YOUR DATA BUCKET]", "gs://[PATH TO YOUR TRAINING DATA]")
# dataset_filename should be contained in your data bucket.
config_params = {}
config_params['CV_folds'] = 1
config_params['cloud_type'] = 'GCP'
config_params['compute'] = 'GPU'
config_params['dataset'] = 'airline'
config_params['dataset_filename'] = 'airline_10000000.orc'
config_params['model_type'] = "RandomForest"
config_params['num_samples'] = 4
config_params['paths'] = paths
config_params['ht_est_range'] = "100,200"
config_params['ht_depth_range'] = "9,17"
config_params['ht_features_range'] = "0.2,0.6"
config_params['ht_experiments'] = 10
model_params = {
'seed': random.random(),
'n_bins': 64
# 'seed': 0
}
accuracy = ax_train(model_params, config_params)
| 0 |
rapidsai_public_repos/cloud-ml-examples/gcp | rapidsai_public_repos/cloud-ml-examples/gcp/docker/example_config.json | {
"trainingInput": {
"args": [
"--train",
"--do-hpo",
"--hpo-num-bins=64",
"--cloud-type=GCP",
"--compute-type=GPU",
"--data-input-path=gs://[YOUR STORAGE BUCKET]",
"--data-output-path=gs://[YOUR STORAGE BUCKET]/training_output",
"--data-name=airline_20000000.orc",
"--model-type=RandomForest"
],
"hyperparameters": {
"enableTrialEarlyStopping": true,
"goal": "MAXIMIZE",
"hyperparameterMetricTag": "hpo_accuracy",
"maxParallelTrials": 8,
"maxTrials": 100,
"maxFailedTrials": 100,
"params": [
{
"maxValue": 200,
"minValue": 100,
"parameterName": "hpo-num-est",
"type": "INTEGER"
},
{
"maxValue": 17,
"minValue": 9,
"parameterName": "hpo-max-depth",
"type": "INTEGER"
},
{
"maxValue": 0.6,
"minValue": 0.2,
"parameterName": "hpo-max-features",
"type": "DOUBLE"
}
]
},
"jobDir": "gs://[YOUR STORAGE BUCKET]/training_output",
"masterConfig": {
"imageUri": "gcr.io/[YOUR GCR PATH]/rapids_training_container:latest",
"acceleratorConfig": {
"count": "1",
"type": "NVIDIA_TESLA_T4"
}
},
"masterType": "n1-standard-8",
"region": "us-west1",
"scaleTier": "CUSTOM"
}
}
| 0 |
rapidsai_public_repos/cloud-ml-examples/gcp | rapidsai_public_repos/cloud-ml-examples/gcp/docker/launch_test.sh | #!/usr/bin/env bash
set -e
set -x
gcloud ai-platform jobs submit training $1 --config ./$2
| 0 |
rapidsai_public_repos/cloud-ml-examples/gcp | rapidsai_public_repos/cloud-ml-examples/gcp/docker/example_config_cpuonly.json | {
"trainingInput": {
"args": [
"--train",
"--do-hpo",
"--hpo-num-bins=64",
"--cloud-type=GCP",
"--compute-type=CPU",
"--data-input-path=gs://[YOUR STORAGE BUCKET]",
"--data-output-path=gs://[YOUR STORAGE BUCKET]/training_output",
"--data-name=airline_20000000.orc",
"--model-type=RandomForest"
],
"hyperparameters": {
"algorithm": "RANDOM_SEARCH",
"enableTrialEarlyStopping": true,
"goal": "MAXIMIZE",
"hyperparameterMetricTag": "hpo_accuracy",
"maxParallelTrials": 10,
"maxTrials": 100,
"maxFailedTrials": 100,
"params": [
{
"maxValue": 600,
"minValue": 100,
"parameterName": "hpo-num-est",
"type": "INTEGER"
},
{
"maxValue": 20,
"minValue": 9,
"parameterName": "hpo-max-depth",
"type": "INTEGER"
},
{
"maxValue": 0.6,
"minValue": 0.2,
"parameterName": "hpo-max-features",
"type": "DOUBLE"
}
]
},
"jobDir": "gs://[YOUR STORAGE BUCKET]/training_output",
"masterConfig": {
"imageUri": "gcr.io/[YOUR PROJECT NAME]/rapids_training_container:latest"
},
"masterType": "n1-standard-8",
"region": "us-west1",
"scaleTier": "CUSTOM"
}
}
| 0 |
rapidsai_public_repos/cloud-ml-examples/gcp/docker | rapidsai_public_repos/cloud-ml-examples/gcp/docker/infrastructure/rapids_lib.py | # os
import sys, os, time, logging
# CPU DS stack
import pandas as pd
import numpy as np
import sklearn
# GPU DS stack [ rapids ]
import gcsfs
# scaling library
import dask
# data ingestion [ CPU ]
from pyarrow import orc as pyarrow_orc
# ML models
from sklearn import ensemble
import xgboost
# data set splits
from sklearn.model_selection import train_test_split as sklearn_train_test_split
# device query
##hack
try:
import cudf, cuml
import pynvml
import cupy
from cuml.model_selection import train_test_split as cuml_train_test_split
import sklearn
from sklearn.model_selection import train_test_split
except:
print("Caught import failures -- probably missing GPU")
# memory query
import psutil
# i/o
import logging, json, pprint
default_sagemaker_paths = {
'base': '/opt/ml',
'code': '/opt/ml/code',
'data': '/opt/ml/input',
'train_data': '/opt/ml/input/data/training',
'hyperparams': '/opt/ml/input/config/hyperparameters.json',
'model': '/opt/ml/model',
'output': '/opt/ml/output',
}
class RapidsCloudML(object):
def __init__(self, cloud_type='AWS',
model_type='XGBoost',
data_type='ORC',
compute_type='single-GPU',
n_workers=-1,
verbose_estimator=False,
CSP_paths=default_sagemaker_paths):
self.CSP_paths = CSP_paths
self.cloud_type = cloud_type
self.model_type = model_type
self.data_type = data_type
self.compute_type = compute_type
self.verbose_estimator = verbose_estimator
self.n_workers = self.parse_compute(n_workers)
self.query_memory()
def _read_orc(self, filename):
if ('CPU' in self.compute_type):
if (filename.startswith('gs://')):
fs = gcsfs.GCSFileSystem()
with fs.open(filename, mode='rb') as file:
dataset = pyarrow_orc.ORCFile(file).read().to_pandas()
else:
with open(filename, mode='rb') as file:
dataset = pyarrow_orc.ORCFile(file).read().to_pandas()
elif ('GPU' in self.compute_type):
dataset = cudf.read_orc(filename)
return dataset
def _read_csv(self, filename, col_labels):
if ('CPU' in self.compute_type):
dataset = pd.read_csv(filename, names=col_labels)
elif ('GPU' in self.compute_type):
dataset = cudf.read_csv(filename, names=col_labels)
return dataset
def load_data(self, filename='dataset.orc', col_labels=None, y_label='ArrDelayBinary'):
target_filename = self.CSP_paths['train_data'] + '/' + filename
self.log_to_file(f'\n> loading dataset from {target_filename}...\n')
with PerfTimer() as ingestion_timer:
if 'ORC' in self.data_type:
dataset = self._read_orc(target_filename)
elif 'CSV' in self.data_type:
dataset = self._read_csv(target_filename, names=col_labels)
self.log_to_file(f'ingestion completed in {ingestion_timer.duration}')
self.log_to_file(f'dataset descriptors: {dataset.shape}\n {dataset.dtypes}\n {dataset.columns}\n')
return dataset, col_labels, y_label, ingestion_timer.duration
def split_data(self, dataset, y_label, train_size=.8, random_state=0, shuffle=True):
"""
split dataset into train and test subset
NOTE: assumes the first column of the dataset is the classification labels
! in the case of sklearn, we manually filter this column in the split call
! in the case of cuml, the filtering happens internally
"""
self.log_to_file('\tsplitting train and test data')
start_time = time.perf_counter()
with PerfTimer() as split_timer:
if 'CPU' in self.compute_type:
X_train, X_test, y_train, y_test = sklearn_train_test_split(dataset.loc[:, dataset.columns != y_label],
dataset[y_label], train_size=train_size,
shuffle=shuffle, random_state=random_state)
elif 'GPU' in self.compute_type:
X_train, X_test, y_train, y_test = cuml_train_test_split(X=dataset, y=y_label, train_size=train_size,
shuffle=shuffle, random_state=random_state)
self.log_to_file(f'\t> split completed in {split_timer.duration}')
return X_train, X_test, y_train, y_test, split_timer.duration
def train_model(self, X_train, y_train, model_params):
self.log_to_file(f'\ttraining {self.model_type} estimator w/ hyper-params')
pprint.pprint(model_params, indent=10)
print(f"model type: {self.model_type}\n compute type: {self.compute_type}\n dataset dtype: {type(X_train)}")
try:
if self.model_type == 'XGBoost':
trained_model, training_time = self.fit_xgboost(X_train, y_train, model_params)
elif self.model_type == 'RandomForest':
trained_model, training_time = self.fit_random_forest(X_train, y_train, model_params)
except Exception as error:
self.log_to_file('!error during model training: ' + str(error))
raise
self.log_to_file(f'\t> finished training in {training_time:.4f} s')
return trained_model, training_time
# train dlmc.xgboost model
def fit_xgboost(self, X_train, y_train, model_params):
with PerfTimer() as train_timer:
train_DMatrix = xgboost.DMatrix(data=X_train, label=y_train)
trained_model = xgboost.train(dtrain=train_DMatrix,
params=model_params,
num_boost_round=model_params['num_boost_round'],
verbose_eval=self.verbose_estimator)
return trained_model, train_timer.duration
# fit_xgboost_multi_GPU ()
# fit_random_forest_multi_GPU ()
# train cuml.random-forest model
def fit_random_forest(self, X_train, y_train, model_params):
if 'CPU' in self.compute_type:
rf_model = sklearn.ensemble.RandomForestClassifier(n_estimators=model_params['n_estimators'],
max_depth=model_params['max_depth'],
max_features=model_params['max_features'],
n_jobs=int(self.n_workers),
verbose=self.verbose_estimator)
elif 'GPU' in self.compute_type:
rf_model = cuml.ensemble.RandomForestClassifier(n_estimators=model_params['n_estimators'],
max_depth=model_params['max_depth'],
n_bins=model_params['n_bins'],
max_features=model_params['max_features'],
verbose=self.verbose_estimator)
with PerfTimer() as train_timer:
trained_model = rf_model.fit(X_train, y_train)
return trained_model, train_timer.duration
def evaluate_test_perf(self, trained_model, X_test, y_test):
self.log_to_file(f'\tinferencing on test set')
with PerfTimer() as inference_timer:
try:
if self.model_type == 'XGBoost':
test_DMatrix = xgboost.DMatrix(data=X_test, label=y_test)
test_accuracy = 1 - float(trained_model.eval(test_DMatrix).split(':')[1])
elif self.model_type == 'RandomForest':
# y_test = cudf.DataFrame({'label': y_test.astype('int32') })
test_accuracy = trained_model.score(X_test, y_test.astype('int32'))
except Exception as error:
self.log_to_file('!error during inference: ' + str(error))
raise
self.log_to_file(f'\t> finished inference in {inference_timer.duration:.4f} s')
return test_accuracy, inference_timer.duration
# TODO: FIL inference [ ? ]
# evaluate_perf_FIL(self, trained_model, X_test, y_test ):
# TODO: global_best_model.save()
def save_best_model(self, global_best_model=None):
pass
# ------------------------------------------------------
# end of data science logic
# ------------------------------------------------------
def parse_compute(self, n_workers=None):
if 'CPU' in self.compute_type or 'GPU' in self.compute_type:
available_devices = self.query_compute()
if n_workers == -1:
n_workers = available_devices
assert (n_workers <= available_devices)
self.log_to_file(f'compute type: {self.compute_type}, n_workers: {n_workers}')
else:
raise Exception('unsupported compute type')
return n_workers
def query_compute(self):
available_devices = None
if 'CPU' in self.compute_type:
available_devices = os.cpu_count()
self.log_to_file(f'detected {available_devices} CPUs')
elif 'GPU' in self.compute_type:
available_devices = cupy.cuda.runtime.getDeviceCount()
self.log_to_file(f'detected {available_devices} GPUs')
return available_devices
# TODO: enumerate all visible GPUs [ ? ]
def query_memory(self):
def print_device_memory(memory, device_ID=-1):
memory_free_GB = np.array(memory.free) / np.array(10e8)
memory_used_GB = np.array(memory.used) / np.array(10e8)
memory_total_GB = np.array(memory.total) / np.array(10e8)
if device_ID != -1:
self.log_to_file(f'device ID = {device_ID}')
self.log_to_file(f'memory free, used, total: {memory_free_GB}, {memory_used_GB}, {memory_total_GB}')
if 'CPU' in self.compute_type:
print_device_memory(psutil.virtual_memory())
elif 'GPU' in self.compute_type:
pynvml.nvmlInit()
for iGPU in range(self.n_workers):
handle = pynvml.nvmlDeviceGetHandleByIndex(iGPU)
print_device_memory(pynvml.nvmlDeviceGetMemoryInfo(handle))
def set_up_logging(self):
logging_path = self.CSP_paths['output'] + '/log.txt'
logging.basicConfig(filename=logging_path,
level=logging.INFO)
def log_to_file(self, text):
logging.info(text)
print(text)
def environment_check(self):
self.check_dirs()
if self.cloud_type == 'AWS':
try:
self.list_files('/opt/ml')
self.log_to_file(os.environ['SM_NUM_GPUS'])
self.log_to_file(os.environ['SM_TRAINING_ENV'])
self.log_to_file(os.environ['SM_CHANNEL_TRAIN'])
self.log_to_file(os.environ['SM_HPS'])
except:
pass
else:
pass
def check_dirs(self):
self.log_to_file('\n> checking for sagemaker paths...\n')
directories_to_check = self.CSP_paths
for iDir, val in directories_to_check.items():
self.log_to_file(f'{val}, exists : {os.path.exists(val)}')
self.log_to_file(f'working directory = {os.getcwd()}')
def list_files(self, startpath):
print(f'\n> listing contents of {startpath}\n')
for root, dirs, files in os.walk(startpath):
level = root.replace(startpath, '').count(os.sep)
indent = ' ' * 4 * (level)
print('{}{}/'.format(indent, os.path.basename(root)))
subindent = ' ' * 4 * (level + 1)
for f in files:
print('{}{}'.format(subindent, f))
# perf_counter = highest available timer resolution
class PerfTimer:
def __init__(self):
self.start = None
self.duration = None
def __enter__(self):
self.start = time.perf_counter()
return self
def __exit__(self, *args):
self.duration = time.perf_counter() - self.start
'''
https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html#sklearn.ensemble.RandomForestClassifier.fit
n_estimators=100,
criterion='gini',
max_depth=None,
min_samples_split=2,
min_samples_leaf=1,
min_weight_fraction_leaf=0.0,
max_features='auto',
max_leaf_nodes=None,
min_impurity_decrease=0.0,
min_impurity_split=None,
bootstrap=True,
oob_score=False,
n_jobs=None,
random_state=None,
verbose=0,
warm_start=False,
class_weight=None,
ccp_alpha=0.0,
max_samples=None
'''
| 0 |
rapidsai_public_repos/cloud-ml-examples/gcp/docker | rapidsai_public_repos/cloud-ml-examples/gcp/docker/infrastructure/entrypoint.py | import argparse
import random
import os
import logging
import hypertune
import json
import sys
from ray import tune
from ray.tune import track
logger = logging.getLogger(tune.__name__)
logger.setLevel(level=logging.CRITICAL)
from rapids_lib import RapidsCloudML
default_sagemaker_paths = {
'base': '/opt/ml',
'code': '/opt/ml/code',
'data': '/opt/ml/input',
'train_data': '/opt/ml/input/data/training',
'hyperparams': '/opt/ml/input/config/hyperparameters.json',
'model': '/opt/ml/model',
'output': '/opt/ml/output',
}
def _train(model_params, config_params):
rcml = RapidsCloudML(cloud_type=config_params['cloud_type'],
model_type=config_params['model_type'],
compute_type=f"single-{args.compute_type}",
CSP_paths=config_params['paths'])
# environment check
rcml.environment_check()
# ingest data [ post pre-processing ]
dataset, col_labels, y_label, ingest_time = rcml.load_data(filename=config_params['dataset_filename'])
rcml.query_memory()
# classifier expects input data to be of type float32
dataset = dataset.astype('float32')
# classification objective requires int32 label for cuml random forest
dataset[y_label] = dataset[y_label].astype('int32')
# ----------------------------------------------------------------------------------------------------
# cross-validation folds
# ----------------------------------------------------------------------------------------------------
global_best_model = None
global_best_test_accuracy = 0
accuracy_per_fold = []
train_time_per_fold = []
infer_time_per_fold = []
split_time_per_fold = []
# optional cross-validation w/ model_params['n_train_folds'] > 1
for i_train_fold in range(config_params['CV_folds']):
rcml.log_to_file(f"\n CV fold {i_train_fold} of {config_params['CV_folds']}\n")
# split data
X_train, X_test, y_train, y_test, split_time = rcml.split_data(dataset=dataset,
y_label=y_label,
random_state=i_train_fold,
shuffle=True)
split_time_per_fold += [round(split_time, 4)]
# train model
trained_model, training_time = rcml.train_model(X_train, y_train, model_params)
train_time_per_fold += [round(training_time, 4)]
# evaluate perf
test_accuracy, infer_time = rcml.evaluate_test_perf(trained_model, X_test, y_test)
accuracy_per_fold += [round(test_accuracy, 4)]
infer_time_per_fold += [round(infer_time, 4)]
# update best model [ assumes maximization of perf metric ]
if test_accuracy > global_best_test_accuracy:
global_best_test_accuracy = test_accuracy
global_best_model = trained_model
rcml.log_to_file(f'\n accuracy per fold : {accuracy_per_fold} \n')
rcml.log_to_file(f'\n train-time per fold : {train_time_per_fold} \n')
rcml.log_to_file(f'\n infer-time per fold : {infer_time_per_fold} \n')
rcml.log_to_file(f'\n split-time per fold : {split_time_per_fold} \n')
return global_best_test_accuracy
def train(model_params, config_params):
"""
Parameters
----------
model_params
config_params
Returns
-------
"""
# ----------------------------------------------------------------------------------------------------
# cross-validation folds
# ----------------------------------------------------------------------------------------------------
global_best_model = None
global_best_test_accuracy = _train(config_params=config_params, model_params=model_params)
return global_best_model, global_best_test_accuracy
def gcp_path_setup(args):
hyperpath = f'{os.environ["RAPIDS_GCP_INSTALL_PATH"]}/hyperparameters.json'
hyperdict = {}
for key, val in args.__dict__.items():
if (key.startswith('hpo')):
new_key = key.replace('hpo-', '')
hyperdict[new_key] = val
with open(hyperpath, 'w') as fpw:
fpw.write(json.dumps(hyperdict, indent=4, sort_keys=True))
paths = {
'train_data': args.data_input_path,
'hyperparams': hyperpath,
'model': f'{args.data_output_path}/model',
'output': f'{args.data_output_path}/output',
}
return paths
def aws_path_setup():
return default_sagemaker_paths
def azure_path_setup():
return {}
def main(args):
paths = {}
if (args.cloud_type.lower() == "gcp"):
paths = gcp_path_setup(args)
elif (args.cloud_type.lower() in ("aws")):
paths = aws_path_setup(args)
elif (args.cloud_type.lower() in ("azure")):
paths = azure_path_setup(args)
config_params = {}
config_params['CV_folds'] = args.cv_folds
config_params['compute'] = args.compute_type
config_params['dataset'] = 'airline'
config_params['dataset_filename'] = args.data_name
config_params['cloud_type'] = args.cloud_type
config_params['model_type'] = args.model_type
config_params['num_samples'] = args.num_samples
config_params['paths'] = paths
config_params['ht_est_range'] = args.ht_est_range
config_params['ht_depth_range'] = args.ht_depth_range
config_params['ht_features_range'] = args.ht_features_range
config_params['ht_experiments'] = args.ht_experiments
if ('RandomForest' in args.model_type):
model_params = {
'max_depth': args.hpo_max_depth,
'max_features': args.hpo_max_features,
'n_bins': args.hpo_num_bins,
'n_estimators': args.hpo_num_est,
'seed': random.random(),
# 'seed': 0
}
elif ('XGBoost' in args.model_type):
model_params = {
'alpha': args.hpo_alpha,
'gamma': args.hpo_gamma,
'lambda': args.hpo_lambda,
'learning_rate': args.hpo_lr,
'max_depth': args.hpo_max_depth,
'num_boost_round': args.hpo_num_boost_round,
'random_state': 0,
'tree_method': 'gpu_hist' if ('GPU' in config_params['compute']) else 'hist'
}
model, accuracy = train(model_params=model_params, config_params=config_params)
if (args.cloud_type.lower() in ("gcp",) and args.do_hpo):
hpt = hypertune.HyperTune()
hpt.report_hyperparameter_tuning_metric(
hyperparameter_metric_tag='hpo_accuracy',
metric_value=accuracy)
if (__name__ in ("__main__",)):
parser = argparse.ArgumentParser()
parser.add_argument('--cloud-type', default='AWS')
parser.add_argument('--compute-type', default='GPU')
parser.add_argument('--data-input-path')
parser.add_argument('--data-output-path')
parser.add_argument('--data-name', default='airline_10000000.orc')
parser.add_argument('--do-hpo', action="store_true", default=False)
parser.add_argument('--epochs', type=int)
parser.add_argument('--hpo-alpha', default=0.0, type=float)
parser.add_argument('--hpo-gamma', default=0.0, type=float)
parser.add_argument('--hpo-lambda', default=1.0, type=float)
parser.add_argument('--hpo-lr', default=0.3, type=float)
parser.add_argument('--hpo-max-depth', default=16, type=int)
parser.add_argument('--hpo-max-features', default=1.0, type=float)
parser.add_argument('--hpo-num-bins', default=64, type=int)
parser.add_argument('--hpo-num-boost-round', default=100, type=int)
parser.add_argument('--hpo-num-est', default=10, type=int)
parser.add_argument('--ht-depth-range', default="9,17", type=str)
parser.add_argument('--ht-est-range', default="100,200", type=str)
parser.add_argument('--ht-features-range', default="0.2,0.6", type=str)
parser.add_argument('--ht-experiments', default=10, type=int)
parser.add_argument('--num-samples', default=4, type=int)
parser.add_argument('--cv-folds', default=1, type=int)
parser.add_argument('--job-dir')
parser.add_argument('--model-type', default="XGBoost", choices=['RandomForest', 'XGBoost'])
parser.add_argument('--train', action="store_true")
args = parser.parse_args()
main(args)
sys.exit(0)
| 0 |
rapidsai_public_repos/cloud-ml-examples | rapidsai_public_repos/cloud-ml-examples/azure/README.md | # RAPIDS on AzureML
These are a few examples to get started on Azure. We'll look at how to set up the environment locally and on Azure to run the notebooks provided.
Sections in README
1. Create an Azure Machine Learning Service Workspace
2. RAPIDS MNMG example using dask-clouprovider
3. RAPIDS Hyperparameter Optimization on AzureML
4. Model Intepretability using GPU SHAP on Azure
5. RAPIDS MNMG with Azure Kubernetes Service (AKS) using Dask Kubernetes
# 1. Create an Azure Machine Learning Service Workspace
### 1(a) Resource Groups and Workspaces
An [Azure Machine Learning service workspace](https://docs.microsoft.com/en-us/azure/machine-learning/concept-workspace) will manage experiments and coordinate storage, databases and computing resources for machine learning applications.
1. First create an [Azure subscription](https://azure.microsoft.com/en-us/free/) or access existing information from the [Azure portal](https://portal.azure.com/).
2. Next you will need to access a [Resource group](https://docs.microsoft.com/en-us/azure/azure-resource-manager/management/overview#resource-groups) or create a new one in Azure portal:
- Sign in to the Azure portal and navigate to Resource groups page by clicking on **Resource groups** in the portal:
- Select one of the available Resource groups or create a new one by clicking on the **Add** button:
- You can also select **+ Create a resource** in the upper-left corner of Azure portal and search for Resource group
Select a a *Subscription* with GPU resources, enter a name for the *Resource group* and select a *Region* with GPU resources. Check these pages for the [List](https://azure.microsoft.com/en-us/global-infrastructure/services/?products=machine-learning-service) of supported regions and [information](https://docs.microsoft.com/en-us/azure/virtual-machines/sizes-gpu) on GPU optimized VM sizes. Pick a region that is closest to your location or contains your data.
3. Next we will create a Machine Learning service workspace: navigate to your Resource groups page and click on the **Add** button, this will take you to the [Azure Marketplace](https://azuremarketplace.microsoft.com/). Use the search bar to find **Machine Learning** or select **AI + Machine Learning** category on the left:
- Click on *Machine Learning* and this will direct you to the page below:
- Enter a unique *Workspace Name* that indentifies your workspace, select your Azure *Subscription*, use an existing *Resource group* in your subscription and select a *Location* with adequate GPU quota.
After entering the information, select **Review + Create**. The deployment success message will appear and and you can view the new workspace by clicking on **Go to resource**.
4. After creating the workspace, download the **config.json** file that includes information about workspace configuration.
This file will be used with [Azure Machine Learning SDK for Python](https://docs.microsoft.com/en-us/python/api/overview/azure/ml/?view=azure-ml-py) in the notebook example to load the workspace and contains a dictionary list with key-values for:
* Workspace name
* Azure region
* Subscription id
* Resource group
# 2. RAPIDS MNMG example using Dask Cloud Provider
The [Azure MNMG notebooks](#) will use [Dask Cloud Provider](https://cloudprovider.dask.org/en/latest/) to run multi-node multi-GPU examples on Azure. For each example, we will make use of [AzureVMCluster](https://cloudprovider.dask.org/en/latest/azure.html#azurevm) function to set-up a cluster and run an example. We have two example notebooks:
- [Random Forest using Dask CloudProvider](./notebooks/Azure-MNMG-RF.ipynb)
- [XGBoost using Dask CloudProvider](./notebooks/Azure-MNMG-XGBoost.ipynb). This notebook additionally demonstrates how to speed up deployment using custom VM images via [`packer`](https://www.packer.io/).
## 2(a) Set up environment on local computer
We recommend using RAPIDS docker image on your local system and using the same image in the notebook so that the libraries can match accurately. You can achieve this using conda environments for RAPIDS too.
For example, in the Random Forest Notebook, we are using `rapidsai/rapidsai:21.06-cuda11.0-runtime-ubuntu18.04-py3.8` docker image, to pull and run this use the following command. The `-v` flag sets the volume you'd like to mount on the docker container. This way, the changes you make within the docker container are present on your local system to. Make sure to change `local/path` to the path which contains this repository.
`docker run --runtime nvidia --rm -it -p 8888:8888 -p 8787:8787 -v /local/path:/docker/path rapidsai/rapidsai:21.06-cuda11.0-runtime-ubuntu18.04-py3.8`
For the XGBoost notebook, we are using the image `rapidsai/rapidsai:cuda11.2-runtime-ubuntu18.04-py3.8`.
## 2(b) Setup Azure environment
We need to setup a Virtual Network and Security Group to run this example. You can use either the command line or the Azure Portal to set these up.
Below, we'll be looking at how you can use command line to set it up. These commands need to be executed within the docker container.
Note: Be sure to set up all the resources in the same region
1. To setup the azure authentication, run `az login`
2. You can make use of the resoruce group you've set up earlier.
3. To create a virtual network - `az network vnet create -g <resource group name> --location <location -n <vnet name> --address-prefix 10.0.0.0/16 --subnet-name <subnet name> --subnet-prefix 10.0.0.0/24`
4. We can now set up the Security group and add a rule for the dask cloud provider run.
```
az network nsg create -g <resource group name> --name <security group name> --location <region>
az network nsg rule create -g <resource group name> --nsg-name <security group name> -n MyNsgRuleWithAsg \
--priority 500 --source-address-prefixes Internet --destination-port-ranges 8786 8787 \
--destination-address-prefixes '*' --access Allow --protocol Tcp --description "Allow Internet to Dask on ports 8786,8787."
```
For more details, visit [Microsoft Azure - dask cloud provider](https://cloudprovider.dask.org/en/latest/azure.html#overview)
5. Once you have set up the resources, start a jupyter notebook on the docker container using the following command
```
jupyter notebook --ip 0.0.0.0 --port 8888 --no-browser --allow-root --NotebookApp.token=''
```
6. Navigate to `Azure-MNMG-RF.ipynb` or `Azure-MNMG-XGBoost.ipynb` under `azure/notebooks`.
7. Update the notebook with the names of resources appropriately and run it.
# 3. RAPIDS Hyperparameter Optimization on AzureML
This example will walk you through how to launch RAPIDS-accelerated hyperparameter optimization jobs on Microsoft Azure ML. Azure ML will train and evaluate models with many different variations of key parameters in order to find the combination that yields the highest accuracy. You'll start by launching a Jupyter notebook locally, which will launch all of the jobs and walk you through the process in more detail.
## 3(a) Set up environment on local computer
Install the [Azure Machine Learning Python SDK](https://docs.microsoft.com/en-us/python/api/overview/azure/ml/install?view=azure-ml-py) (if you are running in your own environment. SDK is already installed in [Azure Notebooks](https://notebooks.azure.com/) or other Microsoft managed environments), this link includes additional instructions to [setup environment](https://docs.microsoft.com/en-us/azure/machine-learning/how-to-configure-environment#local) on your local computer.
After setting up a conda environment, clone the [clould-ml-examples repository](https://github.com/rapidsai/cloud-ml-examples.git) by running the following command in a `local_directory`:
git clone https://github.com/rapidsai/cloud-ml-examples.git
### 3(b) Notebooks and Scripts
Navigate to the azure/notebooks subdirectory. This will include hyperparameter optimizaiton notebooks: HPO-RAPIDS.ipynb and HPO-SKLearn.ipynb. Copy the **config.json** file (that you downloaded after creating a ML workspace) in the directory that contains these notebooks (azure/notebooks). You will load the information from this file in the `Initialize workspace` step of the notebook.
Activate the conda environment, where the Azure ML SDK was installed and launch the Jupyter Notebook server with the following command:
jupyter notebook
Open your web browser, navigate to http://localhost:8888/ and access `HPO-RAPIDS.ipynb` from your local machine. Follow the steps in the notebook for hyperparameter tuning with RAPIDS on GPUs.
# 4. Model Intepretability using GPU SHAP on Azure
1. Follow steps in 1 to [set up a Resource Group and Machine Learning workspace](https://github.com/rapidsai/cloud-ml-examples/blob/main/azure/README.md#1-create-an-azure-machine-learning-service-workspace).
2. Follow steps in [2(a) to set up the local environment](https://github.com/rapidsai/cloud-ml-examples/blob/main/azure/README.md#2a-set-up-environment-on-local-computer) using docker.
3. In the docker container, Clone the cloud-ml-examples repository:
```git
git clone https://github.com/rapidsai/cloud-ml-examples.git
```
4. Navigate to the `azure/notebooks/remote-explanation` and open up `azure-gpu-shap.ipynb`
5. The necessary packages needed are present in the notebook, uncomment and run the appropriate cell.
# 5. RAPIDS MNMG with Azure Kubernetes Service (AKS) using Dask Kubernetes
For detailed instructions of setup and example notebooks to run RAPIDS with Azure Kubernetes Service using Dask Kubernetes, navigate to the `kubernetes` subdirectory.
- Detailed instructions to set up RAPIDS with AKS using Dask Kubernetes is in the markdown file [Detailed_setup_guide.md](./kubernetes/Detailed_setup_guide.md) . Go through this before you try to run any other notebooks.
- Shorter example notebook using Dask + RAPIDS + XGBoost in [MNMG_XGBoost.ipynb](./kubernetes/MNMG_XGBoost.ipynb)
- Full example with performance sweeps over multiple algorithms and larger dataset in [Dask_cuML_Exploration_Full.ipynb](./kubernetes/Dask_cuML_Exploration_Full.ipynb) | 0 |
rapidsai_public_repos/cloud-ml-examples/azure | rapidsai_public_repos/cloud-ml-examples/azure/notebook_setup/README.md | ## **Pack and Deploy Conda Environments for RAPIDS on Microsoft Azure**
This section describes the process required to:
1. Package and deploy a RAPIDS conda environment via helper script
1. Package and deploy a RAPIDS conda environment manually
1. Initialize a RAPIDS conda environment.
1. Package the environment using conda-pack.
1. Unpack into a second second environment with conda-unpack.
1. Unpack an existing conda environment via cloud storage.
### **Package and Deploy Using the Helper Script**
#### Pack Environment
1. `common/code/create_packed_conda_env`
```bash
... processing ...
Packing conda environment
Collecting packages...
Packing environment at '[CONDA ENV]/rapids0.13_py3.7' to 'rapids0.13_py3.7.tar.gz'
[########################################] | 100% Completed | 1min 51.1s
```
#### Unpack Environment on Target System
1. Copy your environment tarball (rapids_py37.tar.gz) to the target system
1. Unpack in desired environment
1. `common/code/create_packed_conda_env --action unpack`
1. Alternatively, the environment can be manually unpacked as
```bash
CONDA_ENV="rapids0.13_py3.7"
TARBALL="$CONDA_ENV.tar.gz"
UNPACK_TO="$CONDA_ENV"
mkdir -p "$UNPACK_TO"
tar -xzf $TARBALL -C "$UNPACK_TO"
source "$UNPACK_TO/bin/activate"
conda-unpack
python -m ipykernel install --user --name $CONDA_ENV
```
### **Package and Deploy Manually**
#### Pack Environment
1. Create a clean conda environment to work from
1. `conda create --name=rapids_env python=3.7`
1. Install conda-pack
1. `conda install -c conda-forge conda-pack`
1. Install RAPIDS
1. Select the package level to install from [RAPIDS.ai](rapids.ai/start.html)
1. Ex. For a full install on Ubuntu 18.04, with CUDA 10.2
```bash
conda install -c rapidsai-nightly -c nvidia -c conda-forge
rapids=0.14 python=3.7 cudatoolkit=10.2
```
1. Pack your environment
1. `conda-pack -n rapids_env -o rapids_py37.tar.gz`
#### Unpack Environment on Target System
1. Copy your environment tarball (rapids_py37.tar.gz) to the target system
1. Extract the tarball to the desired environment
1. Ex. Local
```bash
mkdir -p ./rapids_env
tar -xzf rapids_py37.tar.gz -C ./rapids_env
source ./rapids_env/bin/activate
```
1. Ex. Anaconda
```bash
mkdir -p $HOME/anaconda3/envs/rapids_py37
tar -xzf rapids_py37.tar.gz -C $HOME/anaconda3/envs/rapids_py37
source $HOME/anaconda3/envs/rapids_py37/bin/activate
```
1. Cleanup environment prefixes
1. `conda-unpack`
### **Unpack an Existing Environment Via Cloud Storage**
#### Unpacking on a Target Environment
1. Upload your packed conda environment to Azure's cloud storage
1. Pull and unpack your environment manually or via script as
```bash
AZURE_STORAGE="https://$YOUR_STORAGE_ENDPOINT/$YOUR_BUCKET/rapids_py37.tar.gz"
CONDA_ENV="rapids0.13_py3.7"
TARBALL="$CONDA_ENV.tar.gz"
UNPACK_TO="$CONDA_ENV"
wget -q --show-progress $AZURE_STORAGE
mkdir -p "$UNPACK_TO"
tar -xzf $TARBALL -C "$UNPACK_TO"
source "$UNPACK_TO/bin/activate"
conda-unpack
python -m ipykernel install --user --name $CONDA_ENV
``` | 0 |
rapidsai_public_repos/cloud-ml-examples/azure | rapidsai_public_repos/cloud-ml-examples/azure/code/train_sklearn_RF.py | import argparse
import os
import time
#importing necessary libraries
import numpy as np
import pandas as pd
# import pyarrow
# from pyarrow import orc as pyarrow_orc
import sklearn
from sklearn.ensemble import RandomForestClassifier as sklRF
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
from azureml.core.run import Run
run = Run.get_context()
def main():
start_script = time.time()
parser = argparse.ArgumentParser()
parser.add_argument('--data_dir', type=str, help='location of data')
parser.add_argument('--n_estimators', type=int, default=100, help='Number of trees in RF')
parser.add_argument('--max_depth', type=int, default=16, help='Max depth of each tree')
parser.add_argument('--max_features', type=float, default=1.0, help='Number of features for best split')
args = parser.parse_args()
data_dir = args.data_dir
print('\n---->>>> pandas version <<<<----\n', pd.__version__)
print('\n---->>>> SKLearn version <<<<----\n', sklearn.__version__)
t1 = time.time()
df = pd.read_parquet(os.path.join(data_dir, 'airline_20m_15.parquet'))
# with open( os.path.join(data_dir, 'airline_20000000.orc'), mode='rb') as file:
# df = pyarrow_orc.ORCFile(file).read().to_pandas()
t2 = time.time()
print('\n---->>>> pandas time: {:.2f} <<<<----\n'.format(t2-t1))
X = df[df.columns.difference(['ArrDelay', 'ArrDelayBinary'])]
y = df['ArrDelayBinary'].astype(np.int32)
del df
n_estimators = args.n_estimators
run.log('n_estimators', np.int(args.n_estimators))
max_depth = args.max_depth
run.log('max_depth', np.int(args.max_depth))
max_features = args.max_features
run.log('max_features', np.str(args.max_features))
print('\n---->>>> Training using CPUs <<<<----\n')
# ----------------------------------------------------------------------------------------------------
# cross-validation folds
# ----------------------------------------------------------------------------------------------------
accuracy_per_fold = []; train_time_per_fold = []; infer_time_per_fold = []; trained_model = [];
global_best_model = None; global_best_test_accuracy = 0
traintime = time.time()
# optional cross-validation w/ model_params['n_train_folds'] > 1
for i_train_fold in range(5):
print( f"\n CV fold { i_train_fold } of { 5 }\n" )
# split data
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=i_train_fold, shuffle = True)
# train model
skl_rf = sklRF(n_estimators=n_estimators, max_depth=max_depth, max_features=max_features, n_jobs=-1)
start1 = time.time()
trained_model = skl_rf.fit(X_train, y_train)
training_time = time.time() - start1
train_time_per_fold += [ round( training_time, 4) ]
# evaluate perf
start2 = time.time()
skl_pred = skl_rf.predict(X_test)
infer_time = time.time() - start2
skl_accuracy = accuracy_score(skl_pred, y_test) * 100
accuracy_per_fold += [ round( skl_accuracy, 4) ]
infer_time_per_fold += [ round( infer_time, 4) ]
# update best model [ assumes maximization of perf metric ]
if skl_accuracy > global_best_test_accuracy :
global_best_test_accuracy = skl_accuracy
total_train_inference_time = time.time() - traintime
run.log('Total training inference time', np.float(total_train_inference_time))
run.log('Accuracy', np.float(global_best_test_accuracy))
print( '\n Accuracy :', global_best_test_accuracy)
print( '\n accuracy per fold :', accuracy_per_fold)
print( '\n train-time per fold :', train_time_per_fold)
print( '\n train-time all folds :', sum(train_time_per_fold))
print( '\n infer-time per fold :', infer_time_per_fold)
print( '\n infer-time all folds :', sum(infer_time_per_fold))
end_script = time.time()
print('Total runtime: {:.2f}'.format(end_script-start_script))
run.log('Total runtime', np.float(end_script-start_script))
print('\n Exiting script')
if __name__ == '__main__':
main()
| 0 |
rapidsai_public_repos/cloud-ml-examples/azure | rapidsai_public_repos/cloud-ml-examples/azure/code/train_rapids.py | #
# Copyright (c) 2019-2021, NVIDIA CORPORATION.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
import argparse
import os
import time
import numpy as np
import pandas as pd
import cudf
import cuml
import mlflow
from cuml import RandomForestClassifier as cuRF
from cuml.model_selection import train_test_split
from cuml.metrics.accuracy import accuracy_score
from rapids_csp_azure import RapidsCloudML, PerfTimer
def main():
parser = argparse.ArgumentParser()
parser.add_argument('--data_dir', type=str, help='location of data')
parser.add_argument('--n_estimators', type=int, default=100, help='Number of trees in RF')
parser.add_argument('--max_depth', type=int, default=16, help='Max depth of each tree')
parser.add_argument('--n_bins', type=int, default=8, help='Number of bins used in split point calculation')
parser.add_argument('--max_features', type=float, default=1.0, help='Number of features for best split')
parser.add_argument('--compute', type=str, default='single-GPU', help='set to multi-GPU for algorithms via dask')
parser.add_argument('--cv_folds', type=int, default=5, help='Number of CV fold splits')
args = parser.parse_args()
data_dir = args.data_dir
compute = args.compute
cv_folds = args.cv_folds
n_estimators = args.n_estimators
mlflow.log_param('n_estimators', np.int(args.n_estimators))
max_depth = args.max_depth
mlflow.log_param('max_depth', np.int(args.max_depth))
n_bins = args.n_bins
mlflow.log_param('n_bins', np.int(args.n_bins))
max_features = args.max_features
mlflow.log_param('max_features', np.str(args.max_features))
print('\n---->>>> cuDF version <<<<----\n', cudf.__version__)
print('\n---->>>> cuML version <<<<----\n', cuml.__version__)
azure_ml = RapidsCloudML(cloud_type='Azure', model_type='RandomForest',
data_type='Parquet', compute_type=compute)
print(args.compute)
if compute == 'single-GPU':
dataset, _, y_label, _ = azure_ml.load_data(
filename=os.path.join(data_dir, 'airline_20m.parquet'))
else:
# use parquet files from 'https://airlinedataset.blob.core.windows.net/airline-10years' for multi-GPU training
dataset, _, y_label, _ = azure_ml.load_data(
filename=os.path.join(data_dir, 'part*.parquet'),
col_labels=['Flight_Number_Reporting_Airline',
'Year', 'Quarter', 'Month', 'DayOfWeek',
'DOT_ID_Reporting_Airline', 'OriginCityMarketID',
'DestCityMarketID', 'DepTime', 'DepDelay', 'DepDel15',
'ArrDel15', 'ArrDelay', 'AirTime', 'Distance'],
y_label='ArrDel15')
X = dataset[dataset.columns.difference(['ArrDelay', y_label])]
y = dataset[y_label]
del dataset
print('\n---->>>> Training using GPUs <<<<----\n')
# ----------------------------------------------------------------------------------------------------
# cross-validation folds
# ----------------------------------------------------------------------------------------------------
accuracy_per_fold = []
train_time_per_fold = []
infer_time_per_fold = []
trained_model = []
global_best_test_accuracy = 0
model_params = {
'n_estimators': n_estimators,
'max_depth': max_depth,
'max_features': max_features,
'n_bins': n_bins,
}
# optional cross-validation w/ model_params['n_train_folds'] > 1
for i_train_fold in range(cv_folds):
print(f'\n CV fold { i_train_fold } of { cv_folds }\n')
# split data
X_train, X_test, y_train, y_test, _ = azure_ml.split_data(X, y, random_state=i_train_fold)
# train model
trained_model, training_time = azure_ml.train_model(X_train, y_train, model_params)
train_time_per_fold.append(round(training_time, 4))
# evaluate perf
test_accuracy, infer_time = azure_ml.evaluate_test_perf(trained_model, X_test, y_test)
accuracy_per_fold.append(round(test_accuracy, 4))
infer_time_per_fold.append(round(infer_time, 4))
# update best model [ assumes maximization of perf metric ]
if test_accuracy > global_best_test_accuracy:
global_best_test_accuracy = test_accuracy
mlflow.log_metric('Total training inference time', np.float(training_time + infer_time))
mlflow.log_metric('Accuracy', np.float(global_best_test_accuracy))
print('\n Accuracy :', global_best_test_accuracy)
print('\n accuracy per fold :', accuracy_per_fold)
print('\n train-time per fold :', train_time_per_fold)
print('\n train-time all folds :', sum(train_time_per_fold))
print('\n infer-time per fold :', infer_time_per_fold)
print('\n infer-time all folds :', sum(infer_time_per_fold))
if __name__ == '__main__':
with PerfTimer() as total_script_time:
main()
print('Total runtime: {:.2f}'.format(total_script_time.duration))
mlflow.log_metric('Total runtime', np.float(total_script_time.duration))
print('\n Exiting script')
| 0 |
rapidsai_public_repos/cloud-ml-examples/azure | rapidsai_public_repos/cloud-ml-examples/azure/code/rapids_csp_azure.py | #
# Copyright (c) 2019-2021, NVIDIA CORPORATION.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
import json
import logging
import os
import pprint
import sys
import time
import dask
import numpy as np
import pandas as pd
import psutil
import sklearn
from dask.distributed import Client, wait
from sklearn import ensemble
from sklearn.model_selection import \
train_test_split as sklearn_train_test_split
import cudf
import cuml
import cupy
import dask_cudf
import pynvml
import xgboost
from cuml.dask.common import utils as dask_utils
from cuml.dask.ensemble import RandomForestClassifier as cumlDaskRF
from cuml.metrics.accuracy import accuracy_score
from cuml.model_selection import \
train_test_split as cuml_train_test_split
from dask_cuda import LocalCUDACluster
from dask_ml.model_selection import train_test_split as dask_train_test_split
default_azureml_paths = {
'train_script' : './train_script',
'train_data' : './data_airline',
'output' : './output',
}
class RapidsCloudML(object):
def __init__(self, cloud_type = 'Azure',
model_type = 'RandomForest',
data_type = 'Parquet',
compute_type = 'single-GPU',
verbose_estimator = False,
CSP_paths = default_azureml_paths):
self.CSP_paths = CSP_paths
self.cloud_type = cloud_type
self.model_type = model_type
self.data_type = data_type
self.compute_type = compute_type
self.verbose_estimator = verbose_estimator
self.log_to_file(f'\n> RapidsCloudML\n\tCompute, Data , Model, Cloud types {self.compute_type, self.data_type, self.model_type, self.cloud_type}')
# Setting up client for multi-GPU option
if 'multi' in self.compute_type:
self.log_to_file("\n\tMulti-GPU selected")
# This will use all GPUs on the local host by default
cluster = LocalCUDACluster(threads_per_worker=1)
self.client = Client(cluster)
# Query the client for all connected workers
self.workers = self.client.has_what().keys()
self.n_workers = len(self.workers)
self.log_to_file(f'\n\tClient information {self.client}')
def load_hyperparams(self, model_name = 'XGBoost'):
"""
Selecting model paramters based on the model we select for execution.
Checks if there is a config file present in the path self.CSP_paths['hyperparams'] with
the parameters for the experiment. If not present, it returns the default parameters.
Parameters
----------
model_name : string
Selects which model to set the parameters for. Takes either 'XGBoost' or 'RandomForest'.
Returns
----------
model_params : dict
Loaded model parameters (dict)
"""
self.log_to_file('\n> Loading Hyperparameters')
# Default parameters of the models
if self.model_type == 'XGBoost':
# https://xgboost.readthedocs.io/en/latest/parameter.html
model_params = {
'max_depth': 6,
'num_boost_round': 100,
'learning_rate': 0.3,
'gamma': 0.,
'lambda': 1.,
'alpha': 0.,
'objective':'binary:logistic',
'random_state' : 0
}
elif self.model_type == 'RandomForest':
# https://docs.rapids.ai/api/cuml/stable/ -> cuml.ensemble.RandomForestClassifier
model_params = {
'n_estimators' : 10,
'max_depth' : 10,
'n_bins' : 16,
'max_features': 1.0,
'seed' : 0,
}
hyperparameters = {}
try:
with open(self.CSP_paths['hyperparams'], 'r') as file_handle:
hyperparameters = json.load(file_handle)
for key, value in hyperparameters.items():
model_params[key] = value
pprint.pprint(model_params)
return model_params
except Exception as error:
self.log_to_file(str(error))
return
def load_data(self, filename = 'dataset.orc', col_labels = None, y_label = 'ArrDelayBinary'):
"""
Loading the data into the object from the filename and based on the columns that we are
interested in. Also, generates y_label from 'ArrDelay' column to convert this into a binary
classification problem.
Parameters
----------
filename : string
the path of the dataset to be loaded
col_labels : list of strings
The input columns that we are interested in. None selects all the columns
y_label : string
The column to perform the prediction task in.
Returns
----------
dataset : dataframe (Pandas, cudf or dask-cudf)
Ingested dataset in the format of a dataframe
col_labels : list of strings
The input columns selected
y_label : string
The generated y_label name for binary classification
duration : float
The time it took to execute the function
"""
target_filename = filename
self.log_to_file( f'\n> Loading dataset from {target_filename}')
with PerfTimer() as ingestion_timer:
if 'CPU' in self.compute_type:
# CPU Reading options
self.log_to_file(f'\n\tCPU read')
if self.data_type == 'ORC':
with open( target_filename, mode='rb') as file:
dataset = pyarrow_orc.ORCFile(file).read().to_pandas()
elif self.data_type == 'CSV':
dataset = pd.read_csv( target_filename, names = col_labels )
elif self.data_type == 'Parquet':
if 'single' in self.compute_type:
dataset = pd.read_parquet(target_filename)
elif 'multi' in self.compute_type:
self.log_to_file(f'\n\tReading using dask dataframe')
dataset = dask.dataframe.read_parquet(target_filename, columns = columns)
elif 'GPU' in self.compute_type:
# GPU Reading Option
self.log_to_file(f'\n\tGPU read')
if self.data_type == 'ORC':
dataset = cudf.read_orc(target_filename)
elif self.data_type == 'CSV':
dataset = cudf.read_csv(target_filename, names = col_labels)
elif self.data_type == 'Parquet':
if 'single' in self.compute_type:
dataset = cudf.read_parquet(target_filename)
elif 'multi' in self.compute_type:
self.log_to_file(f'\n\tReading using dask_cudf')
dataset = dask_cudf.read_parquet(target_filename, columns = col_labels)
# cast all columns to float32
for col in dataset.columns:
dataset[col] = dataset[col].astype(np.float32) # needed for random forest
# Adding y_label column if it is not present
if y_label not in dataset.columns:
dataset[y_label] = 1.0 * (
dataset["ArrDelay"] > 10
)
dataset[y_label] = dataset[y_label].astype(np.int32) # Needed for cuml RF
dataset = dataset.fillna(0.0) # Filling the null values. Needed for dask-cudf
self.log_to_file(f'\n\tIngestion completed in {ingestion_timer.duration}')
self.log_to_file(f'\n\tDataset descriptors: {dataset.shape}\n\t{dataset.dtypes}')
return dataset, col_labels, y_label, ingestion_timer.duration
def split_data(self, dataset, y_label, train_size = .8, random_state = 0, shuffle = True):
"""
Splitting data into train and test split, has appropriate imports for different compute modes.
CPU compute - Uses sklearn, we manually filter y_label column in the split call
GPU Compute - Single GPU uses cuml and multi GPU uses dask, both split y_label internally.
Parameters
----------
dataset : dataframe
The dataframe on which we wish to perform the split
y_label : string
The name of the column (not the series itself)
train_size : float
The size for the split. Takes values between 0 to 1.
random_state : int
Useful for running reproducible splits.
shuffle : binary
Specifies if the data must be shuffled before splitting.
Returns
----------
X_train : dataframe
The data to be used for training. Has same type as input dataset.
X_test : dataframe
The data to be used for testing. Has same type as input dataset.
y_train : dataframe
The label to be used for training. Has same type as input dataset.
y_test : dataframe
The label to be used for testing. Has same type as input dataset.
duration : float
The time it took to perform the split
"""
self.log_to_file('\n> Splitting train and test data')
start_time = time.perf_counter()
with PerfTimer() as split_timer:
if 'CPU' in self.compute_type:
X_train, X_test, y_train, y_test = sklearn_train_test_split(dataset.loc[:, dataset.columns != y_label],
dataset[y_label],
train_size = train_size,
shuffle = shuffle,
random_state = random_state)
elif 'GPU' in self.compute_type:
if 'single' in self.compute_type:
X_train, X_test, y_train, y_test = cuml_train_test_split(X = dataset,
y = y_label,
train_size = train_size,
shuffle = shuffle,
random_state = random_state)
elif 'multi' in self.compute_type:
X_train, X_test, y_train, y_test = dask_train_test_split(dataset,
y_label,
train_size = train_size,
shuffle = False, # shuffle not available for dask_cudf yet
random_state = random_state)
self.log_to_file(f'\n\tX_train shape and type{X_train.shape} {type(X_train)}')
self.log_to_file( f'\n\tSplit completed in {split_timer.duration}')
return X_train, X_test, y_train, y_test, split_timer.duration
def train_model(self, X_train, y_train, model_params):
"""
Trains a model with the model_params specified by calling fit_xgboost or
fit_random_forest depending on the model_type.
Parameters
----------
X_train : dataframe
The data for traning
y_train : dataframe
The label to be used for training.
model_params : dict
The model params to use for this training
Returns
----------
trained_model : The object of the trained model either of XGBoost or RandomForest
training_time : float
The time it took to train the model
"""
self.log_to_file(f'\n> Training {self.model_type} estimator w/ hyper-params')
training_time = 0
try:
if self.model_type == 'XGBoost':
trained_model, training_time = self.fit_xgboost(X_train, y_train, model_params)
elif self.model_type == 'RandomForest':
trained_model, training_time = self.fit_random_forest(X_train, y_train, model_params)
except Exception as error:
self.log_to_file('\n\n!error during model training: ' + str(error))
self.log_to_file( f'\n\tFinished training in {training_time:.4f} s')
return trained_model, training_time
def fit_xgboost(self, X_train, y_train, model_params):
"""
Trains a XGBoost model on X_train and y_train with model_params
Parameters and Objects returned are same as trained_model
"""
if 'GPU' in self.compute_type:
model_params.update({'tree_method': 'gpu_hist'})
else:
model_params.update({'tree_method': 'hist'})
with PerfTimer() as train_timer:
if 'single' in self.compute_type:
train_DMatrix = xgboost.DMatrix(data = X_train, label = y_train)
trained_model = xgboost.train(dtrain = train_DMatrix,
params = model_params,
num_boost_round = model_params['num_boost_round'])
elif 'multi' in self.compute_type:
self.log_to_file("\n\tTraining multi-GPU XGBoost")
train_DMatrix = xgboost.dask.DaskDMatrix(self.client, data = X_train, label = y_train)
trained_model = xgboost.dask.train(self.client,
dtrain = train_DMatrix,
params = model_params,
num_boost_round = model_params['num_boost_round'])
return trained_model, train_timer.duration
def fit_random_forest ( self, X_train, y_train, model_params ):
"""
Trains a RandomForest model on X_train and y_train with model_params.
Depending on compute_type, estimators from appropriate packages are used.
CPU - sklearn
Single-GPU - cuml
multi_gpu - cuml.dask
Parameters and Objects returned are same as trained_model
"""
if 'CPU' in self.compute_type:
rf_model = sklearn.ensemble.RandomForestClassifier(n_estimators = model_params['n_estimators'],
max_depth = model_params['max_depth'],
max_features = model_params['max_features'],
n_jobs = int(self.n_workers),
verbose = self.verbose_estimator)
elif 'GPU' in self.compute_type:
if 'single' in self.compute_type:
rf_model = cuml.ensemble.RandomForestClassifier(n_estimators = model_params['n_estimators'],
max_depth = model_params['max_depth'],
n_bins = model_params['n_bins'],
max_features = model_params['max_features'],
verbose = self.verbose_estimator)
elif 'multi' in self.compute_type:
self.log_to_file("\n\tFitting multi-GPU daskRF")
X_train, y_train = dask_utils.persist_across_workers(self.client,
[X_train.fillna(0.0),
y_train.fillna(0.0)],
workers=self.workers)
rf_model = cuml.dask.ensemble.RandomForestClassifier(n_estimators = model_params['n_estimators'],
max_depth = model_params['max_depth'],
n_bins = model_params['n_bins'],
max_features = model_params['max_features'],
verbose = self.verbose_estimator)
with PerfTimer() as train_timer:
try:
trained_model = rf_model.fit( X_train, y_train)
except Exception as error:
self.log_to_file( "\n\n! Error during fit " + str(error))
return trained_model, train_timer.duration
def evaluate_test_perf(self, trained_model, X_test, y_test, threshold=0.5):
"""
Evaluates the model performance on the inference set. For XGBoost we need
to generate a DMatrix and then we can evaluate the model.
For Random Forest, in single GPU case, we can just call .score function.
And multi-GPU Random Forest needs to predict on the model and then compute
the accuracy score.
Parameters
----------
trained_model : The object of the trained model either of XGBoost or RandomForest
X_test : dataframe
The data for testing
y_test : dataframe
The label to be used for testing.
Returns
----------
test_accuracy : float
The accuracy achieved on test set
duration : float
The time it took to evaluate the model
"""
self.log_to_file(f'\n> Inferencing on test set')
test_accuracy = None
with PerfTimer() as inference_timer:
try:
if self.model_type == 'XGBoost':
if 'multi' in self.compute_type:
test_DMatrix = xgboost.dask.DaskDMatrix(self.client, data = X_test, label = y_test)
xgb_pred = xgboost.dask.predict(self.client, trained_model, test_DMatrix).compute()
xgb_pred = (xgb_pred > threshold) * 1.0
test_accuracy = accuracy_score(y_test.compute(), xgb_pred)
elif 'single' in self.compute_type:
test_DMatrix = xgboost.DMatrix(data = X_test, label = y_test)
xgb_pred = trained_model.predict(test_DMatrix)
xgb_pred = (xgb_pred > threshold) * 1.0
test_accuracy = accuracy_score(y_test, xgb_pred)
elif self.model_type == 'RandomForest':
if 'multi' in self.compute_type:
cuml_pred = trained_model.predict(X_test).compute()
self.log_to_file("\n\tPrediction complete")
test_accuracy = accuracy_score(y_test.compute(), cuml_pred, convert_dtype=True)
elif 'single' in self.compute_type:
test_accuracy = trained_model.score( X_test, y_test.astype('int32') )
except Exception as error:
self.log_to_file( '\n\n!error during inference: ' + str(error))
self.log_to_file(f'\n\tFinished inference in {inference_timer.duration:.4f} s')
self.log_to_file(f'\n\tTest-accuracy: {test_accuracy}')
return test_accuracy, inference_timer.duration
def set_up_logging( self ):
"""
Function to set up logging for the object.
"""
logging_path = self.CSP_paths['output'] + '/log.txt'
logging.basicConfig( filename= logging_path,
level=logging.INFO)
def log_to_file ( self, text ):
"""
Logs the text that comes in as input.
"""
logging.info( text )
print(text)
# perf_counter = highest available timer resolution
class PerfTimer:
def __init__(self):
self.start = None
self.duration = None
def __enter__(self):
self.start = time.perf_counter()
return self
def __exit__(self, *args):
self.duration = time.perf_counter() - self.start
| 0 |
rapidsai_public_repos/cloud-ml-examples/azure | rapidsai_public_repos/cloud-ml-examples/azure/notebooks/Train-SKLearn.ipynb | import time
#check core SDK version
import azureml.core
print("SDK version:", azureml.core.VERSION)# data_dir = '../../data_airline_updated'from azureml.core.workspace import Workspace
# if a locally-saved configuration file for the workspace is not available, use the following to load workspace
# ws = Workspace(subscription_id=subscription_id, resource_group=resource_group, workspace_name=workspace_name)
ws = Workspace.from_config()
print('Workspace name: ' + ws.name,
'Azure region: ' + ws.location,
'Subscription id: ' + ws.subscription_id,
'Resource group: ' + ws.resource_group, sep = '\n')
datastore = ws.get_default_datastore()
print("Default datastore's name: {}".format(datastore.name))# datastore.upload(src_dir='../../data_airline_updated', target_path='data_airline', overwrite=False, show_progress=True)path_on_datastore = 'data_airline'
ds_data = datastore.path(path_on_datastore)
print(ds_data)from azureml.core.compute import ComputeTarget, AmlCompute
from azureml.core.compute_target import ComputeTargetException
#choose a name for your cluster
cpu_cluster_name = "cpu-cluster"
if cpu_cluster_name in ws.compute_targets:
cpu_cluster = ws.compute_targets[cpu_cluster_name]
if cpu_cluster and type(cpu_cluster) is AmlCompute:
print('Found compute target. Will use {0} '.format(cpu_cluster_name))
else:
print("creating new cluster")
provisioning_config = AmlCompute.provisioning_configuration(vm_size = 'Standard_DS5_v2', max_nodes = 1)
#create the cluster
cpu_cluster = ComputeTarget.create(ws, cpu_cluster_name, provisioning_config)
#can poll for a minimum number of nodes and for a specific timeout.
#if no min node count is provided it uses the scale settings for the cluster
cpu_cluster.wait_for_completion(show_output=True, min_node_count=None, timeout_in_minutes=20)
#use get_status() to get a detailed status for the current cluster.
print(cpu_cluster.get_status().serialize())import os
project_folder = './train_sklearn'
os.makedirs(project_folder, exist_ok=True)import shutil
shutil.copy('train_sklearn_RF.py', project_folder)from azureml.core import Experiment
experiment_name = 'train_sklearn'
experiment = Experiment(ws, name=experiment_name)from azureml.train.sklearn import SKLearn
script_params = {
'--data_dir': ds_data.as_mount(),
'--n_estimators': 100,
'--max_depth': 8,
'--max_features': 0.6,
}
estimator = SKLearn(source_directory=project_folder,
script_params=script_params,
compute_target=cpu_cluster,
entry_script='train_sklearn_RF.py',
pip_packages=['pyarrow'])run = experiment.submit(estimator)from azureml.widgets import RunDetails
RunDetails(run).show()# run.cancel() | 0 |
rapidsai_public_repos/cloud-ml-examples/azure | rapidsai_public_repos/cloud-ml-examples/azure/notebooks/Azure-MNMG-XGBoost.ipynb | # # Uncomment the following and install some libraries at the beginning.
# # If adlfs is not present, install adlfs to read from Azure data lake.
# ! pip install adlfs
# ! pip install "dask-cloudprovider[azure]" --upgradefrom dask.distributed import Client, wait, get_worker
from dask_cloudprovider.azure import AzureVMCluster
import dask_cudf
from dask_ml.model_selection import train_test_split
from cuml.dask.common import utils as dask_utils
from cuml.metrics import mean_squared_error
from cuml import ForestInference
import cudf
import xgboost as xgb
from datetime import datetime
from dateutil import parser
import numpy as np
from timeit import default_timer as timer
import dask
import json# location = <your chosen location where the resource group and vnet exists>
# resource_group = <your resource group>
# vnet = <vnet in your resource group>
# security_group = <security group in your resource group>
# vm_size = "Standard_NC12s_v3" # or choose a different GPU enabled VM type
docker_image = "rapidsai/rapidsai:cuda11.2-runtime-ubuntu18.04-py3.8"
docker_args = '--shm-size=256m'
worker_class = "dask_cuda.CUDAWorker"
worker_options = {'rmm-managed-memory':True}dask.config.set({"logging.distributed": "info",
"cloudprovider.azure.azurevm.marketplace_plan":{
"publisher": "nvidia",
"name": "ngc-base-version-21-02-2",
"product": "ngc_azure_17_11",
"version": "21.02.2"
}})
vm_image = ""
config = dask.config.get("cloudprovider.azure.azurevm", {})
config# ! az vm image terms accept --urn "nvidia:ngc_azure_17_11:ngc-base-version-21-02-2:21.02.2" --verbosepacker_config = {
"builders": [{
"type": "azure-arm",
"use_azure_cli_auth": True,
"managed_image_resource_group_name": resource_group,
"managed_image_name": <the name of the customized VM image>,
"custom_data_file": "./configs/cloud_init.yaml.j2",
"os_type": "Linux",
"image_publisher": "Canonical",
"image_offer": "UbuntuServer",
"image_sku": "18.04-LTS",
"azure_tags": {
"dept": "RAPIDS-CSP",
"task": "RAPIDS Custom Image deployment"
},
"build_resource_group_name": resource_group,
"vm_size": vm_size
}],
"provisioners": [{
"inline": [
"echo 'Waiting for cloud-init'; while [ ! -f /var/lib/cloud/instance/boot-finished ]; do sleep 1; done; echo 'Done'",
],
"type": "shell"
}]
}
with open("packer_config.json", "w") as fh:
fh.write(json.dumps(packer_config))# # Uncomment the following line and run to create the custom image
# ! packer build packer_config.jsonManagedImageId = <value from the output above> # or the customized VM id if you already have resource id of the customized VM from a previous run. dask.config.set({"cloudprovider.azure.azurevm.vm_image":{}})
config = dask.config.get("cloudprovider.azure.azurevm", {})
print(config)
vm_image = {"id": ManagedImageId}
print(vm_image)%%time
cluster = AzureVMCluster(
location=location,
resource_group=resource_group,
vnet=vnet,
security_group=security_group,
vm_image=vm_image,
vm_size=vm_size,
docker_image=docker_image,
worker_class=worker_class,
n_workers=2,
security=True,
docker_args=docker_args,
worker_options=worker_options,
debug=False,
bootstrap=False, # This is to prevent the cloud init jinja2 script from running in the custom VM.
)client = Client(cluster)
clientdef scale_workers(client, n_workers, n_gpus_per_worker, timeout=300):
import time
client.cluster.scale(n_workers)
m = len(client.has_what().keys())
start = end = time.perf_counter_ns()
while ((m != n_workers*n_gpus_per_worker) and (((end - start) / 1e9) < timeout) ):
time.sleep(5)
m = len(client.has_what().keys())
end = time.perf_counter_ns()
if (((end - start) / 1e9) >= timeout):
raise RuntimeError(f"Failed to rescale cluster in {timeout} sec."
"Try increasing timeout for very large containers, and verify available compute resources.")# # Uncomment if you only have the scheduler with n_workers=0 and want to scale the workers separately.
# %%time
# scale_workers(client, 2, 2, timeout=600)%%time
client.wait_for_workers(2)def pretty_print(scheduler_dict):
print(f"All workers for scheduler id: {scheduler_dict['id']}, address: {scheduler_dict['address']}")
for worker in scheduler_dict['workers']:
print(f"Worker: {worker} , gpu_machines: {scheduler_dict['workers'][worker]['gpu']}")
pretty_print(client.scheduler_info()) # will show some information of the GPUs of the workersdef installAdlfs():
import subprocess
subprocess.run(["pip", "install", "adlfs"])
return "done"
results=client.run(installAdlfs)import math
from math import cos, sin, asin, sqrt, pi
def haversine_distance_kernel(pickup_latitude_r, pickup_longitude_r, dropoff_latitude_r, dropoff_longitude_r, h_distance, radius):
for i, (x_1, y_1, x_2, y_2) in enumerate(zip(pickup_latitude_r, pickup_longitude_r, dropoff_latitude_r, dropoff_longitude_r,)):
x_1 = pi/180 * x_1
y_1 = pi/180 * y_1
x_2 = pi/180 * x_2
y_2 = pi/180 * y_2
dlon = y_2 - y_1
dlat = x_2 - x_1
a = sin(dlat/2)**2 + cos(x_1) * cos(x_2) * sin(dlon/2)**2
c = 2 * asin(sqrt(a))
# radius = 6371 # Radius of earth in kilometers # currently passed as input arguments
h_distance[i] = c * radius
def day_of_the_week_kernel(day, month, year, day_of_week):
for i, (d_1, m_1, y_1) in enumerate(zip(day, month, year)):
if month[i] <3:
shift = month[i]
else:
shift = 0
Y = year[i] - (month[i] < 3)
y = Y - 2000
c = 20
d = day[i]
m = month[i] + shift + 1
day_of_week[i] = (d + math.floor(m*2.6) + y + (y//4) + (c//4) -2*c)%7
def add_features(df):
df['hour'] = df['tpepPickupDateTime'].dt.hour
df['year'] = df['tpepPickupDateTime'].dt.year
df['month'] = df['tpepPickupDateTime'].dt.month
df['day'] = df['tpepPickupDateTime'].dt.day
df['diff'] = (df['tpepPickupDateTime'] - df['tpepPickupDateTime']).dt.seconds #convert difference between pickup and dropoff into seconds
df['pickup_latitude_r'] = df['startLat']//.01*.01
df['pickup_longitude_r'] = df['startLon']//.01*.01
df['dropoff_latitude_r'] = df['endLat']//.01*.01
df['dropoff_longitude_r'] = df['endLon']//.01*.01
df = df.drop('tpepDropoffDateTime', axis=1)
df = df.drop('tpepPickupDateTime', axis =1)
df = df.apply_rows(haversine_distance_kernel,
incols=['pickup_latitude_r', 'pickup_longitude_r', 'dropoff_latitude_r', 'dropoff_longitude_r'],
outcols=dict(h_distance=np.float32),
kwargs=dict(radius=6371))
df = df.apply_rows(day_of_the_week_kernel,
incols=['day', 'month', 'year'],
outcols=dict(day_of_week=np.float32),
kwargs=dict())
df['is_weekend'] = (df['day_of_week']<2)
return dfdef persist_train_infer_split(client, df, response_dtype, response_id, infer_frac=1.0, random_state=42, shuffle=True):
workers = client.has_what().keys()
X, y = df.drop([response_id], axis=1), df[response_id].astype('float32')
infer_frac = max(0, min(infer_frac, 1.0))
X_train, X_infer, y_train, y_infer = train_test_split(X, y, shuffle=True, random_state=random_state, test_size=infer_frac)
with dask.annotate(workers=set(workers)):
X_train, y_train = client.persist(
collections=[X_train, y_train])
if (infer_frac != 1.0):
with dask.annotate(workers=set(workers)):
X_infer, y_infer = client.persist(
collections=[X_infer, y_infer])
wait([X_train, y_train, X_infer, y_infer])
else:
X_infer = X_train
y_infer = y_train
wait([X_train, y_train])
return X_train, y_train, X_infer, y_infer
def clean(df_part, must_haves):
"""
This function performs the various clean up tasks for the data
and returns the cleaned dataframe.
"""
# iterate through columns in this df partition
for col in df_part.columns:
# drop anything not in our expected list
if col not in must_haves:
df_part = df_part.drop(col, axis=1)
continue
# fixes datetime error found by Ty Mckercher and fixed by Paul Mahler
if df_part[col].dtype == 'object' and col in ['tpepPickupDateTime', 'tpepDropoffDateTime']:
df_part[col] = df_part[col].astype('datetime64[ms]')
continue
# if column was read as a string, recast as float
if df_part[col].dtype == 'object':
df_part[col] = df_part[col].str.fillna('-1')
df_part[col] = df_part[col].astype('float32')
else:
# downcast from 64bit to 32bit types
# Tesla T4 are faster on 32bit ops
if 'int' in str(df_part[col].dtype):
df_part[col] = df_part[col].astype('int32')
if 'float' in str(df_part[col].dtype):
df_part[col] = df_part[col].astype('float32')
df_part[col] = df_part[col].fillna(-1)
return df_part
def taxi_data_loader(client, adlsaccount, adlspath, response_dtype=np.float32, infer_frac=1.0, random_state=0):
#create a list of columns & dtypes the df must have
must_haves = {
'tpepPickupDateTime': 'datetime64[ms]',
'tpepDropoffDateTime': 'datetime64[ms]',
'passengerCount': 'int32',
'tripDistance': 'float32',
'startLon': 'float32',
'startLat': 'float32',
'rateCodeId': 'int32',
'endLon': 'float32',
'endLat': 'float32',
'fareAmount': 'float32'
}
workers = client.has_what().keys()
response_id = 'fareAmount'
storage_options = {'account_name': adlsaccount}
taxi_data = dask_cudf.read_parquet(adlspath, storage_options=storage_options, chunksize=25e6, npartitions=len(workers))
taxi_data = clean(taxi_data, must_haves)
taxi_data = taxi_data.map_partitions(add_features)
# Drop NaN values and convert to float32
taxi_data = taxi_data.dropna()
fields = ['passengerCount', 'tripDistance', 'startLon', 'startLat', 'rateCodeId',
'endLon', 'endLat', 'fareAmount', 'diff', 'h_distance', 'day_of_week', 'is_weekend']
taxi_data = taxi_data.astype("float32")
taxi_data = taxi_data[fields]
taxi_data = taxi_data.reset_index()
return persist_train_infer_split(client, taxi_data, response_dtype, response_id, infer_frac, random_state)tic = timer()
X_train, y_train, X_infer, y_infer = taxi_data_loader(client,
adlsaccount="azureopendatastorage",
adlspath="az://nyctlc/yellow/puYear=2014/puMonth=1*/*.parquet",
infer_frac=0.1, random_state=42)
toc = timer()
print(f"Wall clock time taken for ETL and persisting : {toc-tic} s")X_train.shape[0].compute()X_train.head()params = {
'learning_rate': 0.15,
'max_depth': 8,
'objective': 'reg:squarederror',
'subsample': 0.7,
'colsample_bytree': 0.7,
'min_child_weight': 1,
'gamma': 1,
'silent': True,
'verbose_eval': True,
'booster' : 'gbtree', # 'gblinear' not implemented in dask
'debug_synchronize': True,
'eval_metric': 'rmse',
'tree_method':'gpu_hist',
'num_boost_rounds': 100,
}data_train = xgb.dask.DaskDMatrix(client, X_train, y_train)
tic = timer()
xgboost_output = xgb.dask.train(client, params,data_train,
num_boost_round=params['num_boost_rounds'])
xgb_gpu_model = xgboost_output['booster']
toc = timer()
print(f"Wall clock time taken for this cell : {toc-tic} s")model_filename = 'trained-model_nyctaxi.xgb'
xgb_gpu_model.save_model(model_filename)_y_test = y_infer.compute()
wait(_y_test)d_test = xgb.dask.DaskDMatrix(client, X_infer)
tic = timer()
y_pred = xgb.dask.predict(client, xgb_gpu_model, d_test)
y_pred= y_pred.compute()
wait(y_pred)
toc = timer()
print(f"Wall clock time taken for xgb.dask.predict : {toc-tic} s")tic = timer()
y_pred = xgb.dask.inplace_predict(client, xgb_gpu_model, X_infer)
y_pred = y_pred.compute()
wait(y_pred)
toc = timer()
print(f"Wall clock time taken for inplace inference : {toc-tic} s")tic = timer()
print("Calculating MSE")
score = mean_squared_error(y_pred, _y_test)
print("Workflow Complete - RMSE: ", np.sqrt(score))
toc = timer()
print(f"Wall clock time taken for this cell : {toc-tic} s")from cuml import ForestInference
from dask.distributed import get_workerworkers = client.has_what().keys()
print(workers)
n_workers = len(workers)
n_partitions = n_workersdef unzipFile(zipname):
worker = get_worker()
import zipfile
import os
with zipfile.ZipFile(os.path.join(worker.local_directory, zipname)) as zf:
zf.extractall(worker.local_directory)
def checkOrMakeLocalDir():
worker = get_worker()
import os
if not os.path.exists(worker.local_directory):
os.makedirs(worker.local_directory)
def workerModelInit(model_file):
# this function will run in each worker and initialize the worker
import os
worker = get_worker()
worker.data["fil_model"] = ForestInference.load(filename=os.path.join(worker.local_directory, model_file),model_type='xgboost')
def predict(input_df):
# this function will run in each worker and predict
worker = get_worker()
return worker.data["fil_model"].predict(input_df)
def persistModelonWorkers(client, zip_file_name, model_file_name):
import zipfile
zf = zipfile.ZipFile(zip_file_name, mode='w')
zf.write(f"./{model_file_name}")
zf.close()
# check to see if local directory present in workers
# if not present make it
fut = client.run(checkOrMakeLocalDir)
wait(fut)
# upload the zip file in workers
fut = client.upload_file(f"./{zip_file_name}")
wait(fut)
# unzip file in the workers
fut = client.run(unzipFile, zip_file_name)
wait(fut)
# load model using FIL in workers
fut = client.run(workerModelInit, model_file_name)
wait(fut)
%%time
persistModelonWorkers(client, "zipfile_write.zip", "trained-model_nyctaxi.xgb")tic = timer()
predictions = X_infer.map_partitions(predict, meta="float") # this is like MPI reduce
y_pred = predictions.compute()
wait(y_pred)
toc = timer()
print(f"Wall clock time taken for this cell : {toc-tic} s")rows_csv = X_infer.iloc[:,0].shape[0].compute()
print(f"It took {toc-tic} seconds to predict on {rows_csv} rows using FIL distributedly on each worker")tic = timer()
score = mean_squared_error(y_pred, _y_test)
toc = timer()
print("Final - RMSE: ", np.sqrt(score))client.close()
cluster.close() | 0 |
rapidsai_public_repos/cloud-ml-examples/azure | rapidsai_public_repos/cloud-ml-examples/azure/notebooks/Train-RAPIDS.ipynb | # verify installation and check Azure ML SDK version
import azureml.core
print('SDK version:', azureml.core.VERSION)from azureml.core import Workspace
# if a locally-saved configuration file for the workspace is not available, use the following to load workspace
# ws = Workspace(subscription_id=subscription_id, resource_group=resource_group, workspace_name=workspace_name)
ws = Workspace.from_config()
print('Workspace name: ' + ws.name,
'Azure region: ' + ws.location,
'Subscription id: ' + ws.subscription_id,
'Resource group: ' + ws.resource_group, sep = '\n')from azureml.core import Dataset
ds = Dataset.File.from_files("https://airlinedataset.blob.core.windows.net/airline-20m/*")from azureml.core.compute import ComputeTarget, AmlCompute
from azureml.core.compute_target import ComputeTargetException
# choose a name for your cluster
gpu_cluster_name = 'gpu-cluster'
if gpu_cluster_name in ws.compute_targets:
gpu_cluster = ws.compute_targets[gpu_cluster_name]
if gpu_cluster and type(gpu_cluster) is AmlCompute:
print('Found compute target. Will use {0} '.format(gpu_cluster_name))
else:
print('creating new cluster')
# m_size parameter below could be modified to one of the RAPIDS-supported VM types
provisioning_config = AmlCompute.provisioning_configuration(vm_size = 'Standard_NC6s_v3', max_nodes = 1, idle_seconds_before_scaledown = 300, vm_priority = "lowpriority")
# Use VM types with more than one GPU for multi-GPU option, e.g. Standard_NC12s_v3
# create the cluster
gpu_cluster = ComputeTarget.create(ws, gpu_cluster_name, provisioning_config)
# can poll for a minimum number of nodes and for a specific timeout
# if no min node count is provided it uses the scale settings for the cluster
gpu_cluster.wait_for_completion(show_output=True, min_node_count=None, timeout_in_minutes=20)
# use get_status() to get a detailed status for the current cluster
print(gpu_cluster.get_status().serialize())import os
project_folder = './train_rapids'
os.makedirs(project_folder, exist_ok=True)notebook_path = os.path.realpath('__file__'+'/../../code')
rapids_script = os.path.join(notebook_path, 'train_rapids.py')
azure_script = os.path.join(notebook_path, 'rapids_csp_azure.py')import shutil
shutil.copy(rapids_script, project_folder)
shutil.copy(azure_script, project_folder)from azureml.core import Experiment
experiment_name = 'train_rapids'from azureml.core import Environment
#environment file
environment_file = "Dockerfile"
environment_name = "rapids"
env = Environment(environment_name)
env.docker.enabled = True
env.docker.base_image = None
env.docker.base_dockerfile = environment_file
env.python.user_managed_dependencies = Truefrom azureml.core import ScriptRunConfig
script_params = [
'--data_dir', ds.as_mount(),
'--n_estimators', 100,
'--max_depth', 8,
'--n_bins', 8,
'--max_features', 0.6,
]
src = ScriptRunConfig(source_directory=project_folder,
arguments=script_params,
compute_target=gpu_cluster,
script='train_rapids.py',
environment=env)run = Experiment(ws, experiment_name).submit(src)runfrom azureml.widgets import RunDetails
RunDetails(run).show()# run.cancel()# gpu_cluster.delete() | 0 |
rapidsai_public_repos/cloud-ml-examples/azure | rapidsai_public_repos/cloud-ml-examples/azure/notebooks/HPO-RAPIDS.ipynb | # verify installation and check Azure ML SDK version
import azureml.core
print('SDK version:', azureml.core.VERSION)from azureml.core import Dataset
airline_ds = Dataset.File.from_files("https://airlinedataset.blob.core.windows.net/airline-20m/*")
# larger dataset (10 years of airline data) is also available for multi-GPU option
# airline_ds = Dataset.File.from_files('https://airlinedataset.blob.core.windows.net/airline-10years/*')from azureml.core import Workspace
# if a locally-saved configuration file for the workspace is not available, use the following to load workspace
# ws = Workspace(subscription_id=subscription_id, resource_group=resource_group, workspace_name=workspace_name)
ws = Workspace.from_config()
print('Workspace name: ' + ws.name,
'Azure region: ' + ws.location,
'Subscription id: ' + ws.subscription_id,
'Resource group: ' + ws.resource_group, sep = '\n')from azureml.core.compute import ComputeTarget, AmlCompute
from azureml.core.compute_target import ComputeTargetException
# choose a name for your cluster
gpu_cluster_name = 'gpu-cluster'
if gpu_cluster_name in ws.compute_targets:
gpu_cluster = ws.compute_targets[gpu_cluster_name]
if gpu_cluster and type(gpu_cluster) is AmlCompute:
print('Found compute target. Will use {0} '.format(gpu_cluster_name))
else:
print('creating new cluster')
# m_size parameter below could be modified to one of the RAPIDS-supported VM types
provisioning_config = AmlCompute.provisioning_configuration(vm_size = 'Standard_NC6s_v3', max_nodes = 5, idle_seconds_before_scaledown = 300)
# Use VM types with more than one GPU for multi-GPU option, e.g. Standard_NC12s_v3
# create the cluster
gpu_cluster = ComputeTarget.create(ws, gpu_cluster_name, provisioning_config)
# can poll for a minimum number of nodes and for a specific timeout
# if no min node count is provided it uses the scale settings for the cluster
gpu_cluster.wait_for_completion(show_output=True, min_node_count=None, timeout_in_minutes=20)
# use get_status() to get a detailed status for the current cluster
print(gpu_cluster.get_status().serialize())import os
project_folder = './train_rapids'
os.makedirs(project_folder, exist_ok=True)notebook_path = os.path.realpath('__file__'+'/../../code')
rapids_script = os.path.join(notebook_path, 'train_rapids.py')
azure_script = os.path.join(notebook_path, 'rapids_csp_azure.py')import shutil
shutil.copy(rapids_script, project_folder)
shutil.copy(azure_script, project_folder)from azureml.core import Experiment
experiment_name = 'train_rapids'from azureml.core import Environment
from azureml.core.runconfig import DockerConfiguration
environment_name = 'rapids_hpo'
env = Environment(environment_name)
# enable docker
docker_config = DockerConfiguration(use_docker=True)
# rapids-cloud-ml image is available in Docker Hub
env.docker.base_image = 'rapidsai/rapidsai-cloud-ml:latest'
# use rapids environment in the container, don't build a new conda environment
env.python.user_managed_dependencies = Truefrom azureml.core import ScriptRunConfig
arguments = [
'--data_dir', airline_ds.as_mount(),
'--n_bins', 32,
'--compute', 'single-GPU', # set to multi-GPU for algorithms via Dask
'--cv_folds', 5,
]
src = ScriptRunConfig(source_directory=project_folder,
arguments=arguments,
compute_target=gpu_cluster,
script='train_rapids.py',
environment=env,
docker_runtime_config=docker_config)from azureml.train.hyperdrive.runconfig import HyperDriveConfig
from azureml.train.hyperdrive.sampling import RandomParameterSampling
from azureml.train.hyperdrive.run import PrimaryMetricGoal
from azureml.train.hyperdrive.parameter_expressions import choice, loguniform, uniform
param_sampling = RandomParameterSampling( {
'--n_estimators': choice(range(50, 500)),
'--max_depth': choice(range(5, 19)),
'--max_features': uniform(0.2, 1.0)
}
)
hyperdrive_run_config = HyperDriveConfig(run_config=src,
hyperparameter_sampling=param_sampling,
primary_metric_name='Accuracy',
primary_metric_goal=PrimaryMetricGoal.MAXIMIZE,
max_total_runs=10,
max_concurrent_runs=5)# start the HyperDrive run
run = Experiment(ws, experiment_name).submit(hyperdrive_run_config)
runfrom azureml.widgets import RunDetails
RunDetails(run).show()run.wait_for_completion(show_output=True)# run.cancel()best_run = hyperdrive_run.get_best_run_by_primary_metric()
print(best_run.get_details()['runDefinition']['arguments'])print(best_run.get_file_names())# model = best_run.register_model(model_name='train-rapids', model_path='outputs/model-rapids.joblib')# gpu_cluster.delete() | 0 |
rapidsai_public_repos/cloud-ml-examples/azure | rapidsai_public_repos/cloud-ml-examples/azure/notebooks/Dockerfile | FROM rapidsai/rapidsai-core:21.06-cuda11.0-base-ubuntu18.04-py3.8
RUN apt-get update && \
apt-get install -y fuse && \
source activate rapids && \
pip install azureml-mlflow && \
pip install azureml-dataprep && \
pip install dask-ml
| 0 |
rapidsai_public_repos/cloud-ml-examples/azure | rapidsai_public_repos/cloud-ml-examples/azure/notebooks/HPO-SKLearn.ipynb | # verify installation and check Azure ML SDK version
import azureml.core
print('SDK version:', azureml.core.VERSION)from azureml.core.dataset import Dataset
airline_ds = Dataset.File.from_files('https://airlinedataset.blob.core.windows.net/airline-20m/*')
# larger dataset (10 years of airline data) is also available for multi-GPU option
# airline_ds = Dataset.File.from_files('https://airlinedataset.blob.core.windows.net/airline-10years/*')# download the dataset as local files
airline_ds.download(target_path='/local/path')from azureml.core.workspace import Workspace
# if a locally-saved configuration file for the workspace is not available, use the following to load workspace
# ws = Workspace(subscription_id=subscription_id, resource_group=resource_group, workspace_name=workspace_name)
ws = Workspace.from_config()
print('Workspace name: ' + ws.name,
'Azure region: ' + ws.location,
'Subscription id: ' + ws.subscription_id,
'Resource group: ' + ws.resource_group, sep = '\n')
datastore = ws.get_default_datastore()
print("Default datastore's name: {}".format(datastore.name))path_on_datastore = 'data_airline'
datastore.upload(src_dir='/add/local/path', target_path=path_on_datastore, overwrite=False, show_progress=True)ds_data = datastore.path(path_on_datastore)
print(ds_data)from azureml.core.compute import ComputeTarget, AmlCompute
from azureml.core.compute_target import ComputeTargetException
# choose a name for your cluster
cpu_cluster_name = "cpu-cluster"
if cpu_cluster_name in ws.compute_targets:
cpu_cluster = ws.compute_targets[cpu_cluster_name]
if cpu_cluster and type(cpu_cluster) is AmlCompute:
print('Found compute target. Will use {0} '.format(cpu_cluster_name))
else:
print("creating new cluster")
provisioning_config = AmlCompute.provisioning_configuration(vm_size = 'Standard_DS5_v2', max_nodes = 10, idle_seconds_before_scaledown = 300)
# create the cluster
cpu_cluster = ComputeTarget.create(ws, cpu_cluster_name, provisioning_config)
# can poll for a minimum number of nodes and for a specific timeout.
# if no min node count is provided it uses the scale settings for the cluster
cpu_cluster.wait_for_completion(show_output=True, min_node_count=None, timeout_in_minutes=20)
# use get_status() to get a detailed status for the current cluster.
print(cpu_cluster.get_status().serialize())import os
project_folder = './train_sklearn'
os.makedirs(project_folder, exist_ok=True)import shutil
shutil.copy('../code/train_sklearn_RF.py', project_folder)from azureml.core import Experiment
experiment_name = 'train_sklearn'
experiment = Experiment(ws, name=experiment_name)from azureml.train.sklearn import SKLearn
script_params = {
'--data_dir': ds_data.as_mount(),
}
estimator = SKLearn(source_directory=project_folder,
script_params=script_params,
compute_target=cpu_cluster,
entry_script='train_sklearn_RF.py',
pip_packages=['pyarrow'])from azureml.train.hyperdrive.runconfig import HyperDriveConfig
from azureml.train.hyperdrive.sampling import RandomParameterSampling
from azureml.train.hyperdrive.run import PrimaryMetricGoal
from azureml.train.hyperdrive.parameter_expressions import choice, loguniform, uniform
param_sampling = RandomParameterSampling( {
'--n_estimators': choice(range(50, 500)),
'--max_depth': choice(range(5, 19)),
'--max_features': uniform(0.2, 1.0)
}
)
hyperdrive_run_config = HyperDriveConfig(estimator=estimator,
hyperparameter_sampling=param_sampling,
primary_metric_name='Accuracy',
primary_metric_goal=PrimaryMetricGoal.MAXIMIZE,
max_total_runs=100,
max_concurrent_runs=10)# start the HyperDrive run
hyperdrive_run = experiment.submit(hyperdrive_run_config)from azureml.widgets import RunDetails
RunDetails(hyperdrive_run).show()# hyperdrive_run.wait_for_completion(show_output=True)# hyperdrive_run.cancel()best_run = hyperdrive_run.get_best_run_by_primary_metric()
print(best_run.get_details()['runDefinition']['arguments'])print(best_run.get_file_names())# model = best_run.register_model(model_name='train-sklearn', model_path='outputs/model-sklearn.joblib')# delete the cluster
# gpu_cluster.delete() | 0 |
rapidsai_public_repos/cloud-ml-examples/azure | rapidsai_public_repos/cloud-ml-examples/azure/notebooks/Azure-MNMG-RF.ipynb | # !pip install "dask-cloudprovider[azure]"
# !pip install azureml-core
# # Run the statements below one after the other in order.
# !pip install azureml-opendatasets
# !pip install --upgrade pandasimport math
from datetime import datetime
from math import asin, cos, pi, sin, sqrt
import cudf
import dask
import dask_cudf
import numpy as np
# This is a package in preview.
from azureml.opendatasets import NycTlcYellow
from cuml.dask.common import utils as dask_utils
from cuml.dask.ensemble import RandomForestRegressor
from cuml.metrics import mean_squared_error
from dask.distributed import Client, wait
from dask_cloudprovider.azure import AzureVMCluster
from dask_ml.model_selection import train_test_split
from dateutil import parserlocation = "SOUTH CENTRAL US"
resource_group = "RAPIDS-MNMG"
vnet = "dask-vnet"
security_group = "dask-nsg"
vm_size = "Standard_NC12s_v3"
docker_image = "rapidsai/rapidsai:21.06-cuda11.0-runtime-ubuntu18.04-py3.8"
docker_args = '--shm-size=256m'
worker_class = "dask_cuda.CUDAWorker"
worker_options = {'rmm-managed-memory':True}
n_workers = 2dask.config.set({"logging.distributed": "info",
"cloudprovider.azure.azurevm.marketplace_plan": {
"publisher": "nvidia",
"name": "ngc-base-version-21-02-2",
"product": "ngc_azure_17_11",
"version": "21.02.2"
}})
vm_image = ""
config = dask.config.get("cloudprovider.azure.azurevm", {})
config%%time
cluster = AzureVMCluster(
location=location,
resource_group=resource_group,
vnet=vnet,
security_group=security_group,
vm_image=vm_image,
vm_size=vm_size,
docker_image=docker_image,
worker_class=worker_class,
n_workers=n_workers,
security=True,
docker_args=docker_args,
worker_options=worker_options,
debug=False,
bootstrap=False, # This is to prevent the cloud init jinja2 script from running in the custom VM.
)#create a list of columns & dtypes the df must have
must_haves = {
'tpepPickupDateTime': 'datetime64[ms]',
'tpepDropoffDateTime': 'datetime64[ms]',
'passengerCount': 'int32',
'tripDistance': 'float32',
'startLon': 'float32',
'startLat': 'float32',
'rateCodeId': 'int32',
'endLon': 'float32',
'endLat': 'float32',
'fareAmount': 'float32'
}def clean(df_part, must_haves):
"""
This function performs the various clean up tasks for the data
and returns the cleaned dataframe.
"""
# iterate through columns in this df partition
for col in df_part.columns:
# drop anything not in our expected list
if col not in must_haves:
df_part = df_part.drop(col, axis=1)
continue
# fixes datetime error found by Ty Mckercher and fixed by Paul Mahler
if df_part[col].dtype == 'object' and col in ['tpepPickupDateTime', 'tpepDropoffDateTime']:
df_part[col] = df_part[col].astype('datetime64[ms]')
continue
# if column was read as a string, recast as float
if df_part[col].dtype == 'object':
df_part[col] = df_part[col].str.fillna('-1')
df_part[col] = df_part[col].astype('float32')
else:
# downcast from 64bit to 32bit types
# Tesla T4 are faster on 32bit ops
if 'int' in str(df_part[col].dtype):
df_part[col] = df_part[col].astype('int32')
if 'float' in str(df_part[col].dtype):
df_part[col] = df_part[col].astype('float32')
df_part[col] = df_part[col].fillna(-1)
return df_partdef haversine_distance_kernel(startLat, startLon, endLat, endLon, h_distance):
for i, (x_1, y_1, x_2, y_2) in enumerate(zip(startLat, startLon, endLat, endLon,)):
x_1 = pi/180 * x_1
y_1 = pi/180 * y_1
x_2 = pi/180 * x_2
y_2 = pi/180 * y_2
dlon = y_2 - y_1
dlat = x_2 - x_1
a = sin(dlat/2)**2 + cos(x_1) * cos(x_2) * sin(dlon/2)**2
c = 2 * asin(sqrt(a))
r = 6371 # Radius of earth in kilometers
h_distance[i] = c * r
def day_of_the_week_kernel(day, month, year, day_of_week):
for i, (d_1, m_1, y_1) in enumerate(zip(day, month, year)):
if month[i] <3:
shift = month[i]
else:
shift = 0
Y = year[i] - (month[i] < 3)
y = Y - 2000
c = 20
d = day[i]
m = month[i] + shift + 1
day_of_week[i] = (d + math.floor(m*2.6) + y + (y//4) + (c//4) -2*c)%7
def add_features(df):
df['hour'] = df['tpepPickupDateTime'].dt.hour
df['year'] = df['tpepPickupDateTime'].dt.year
df['month'] = df['tpepPickupDateTime'].dt.month
df['day'] = df['tpepPickupDateTime'].dt.day
df['diff'] = df['tpepDropoffDateTime'].astype('int32') - df['tpepPickupDateTime'].astype('int32')
df['pickup_latitude_r'] = df['startLat']//.01*.01
df['pickup_longitude_r'] = df['startLon']//.01*.01
df['dropoff_latitude_r'] = df['endLat']//.01*.01
df['dropoff_longitude_r'] = df['endLon']//.01*.01
df = df.drop('tpepDropoffDateTime', axis=1)
df = df.drop('tpepPickupDateTime', axis =1)
df = df.apply_rows(haversine_distance_kernel,
incols=['startLat', 'startLon', 'endLat', 'endLon'],
outcols=dict(h_distance=np.float32),
kwargs=dict())
df = df.apply_rows(day_of_the_week_kernel,
incols=['day', 'month', 'year'],
outcols=dict(day_of_week=np.float32),
kwargs=dict())
df['is_weekend'] = (df['day_of_week']<2)
return df
def scale_workers(client, n_workers, n_gpus_per_worker, timeout=300):
import time
client.cluster.scale(n_workers)
m = len(client.has_what().keys())
start = end = time.perf_counter_ns()
while ((m != n_workers*n_gpus_per_worker) and (((end - start) / 1e9) < timeout) ):
time.sleep(5)
m = len(client.has_what().keys())
end = time.perf_counter_ns()
if (((end - start) / 1e9) >= timeout):
raise RuntimeError(f"Failed to rescale cluster in {timeout} sec."
"Try increasing timeout for very large containers, and verify available compute resources.")client = Client(cluster)
# Scale workers and wait for workers to be up and running
# Number of GPUs per node for the VM we've spun up is 2
scale_workers(client, n_workers, 2, timeout=600) # Run this just once per cluster
client.wait_for_workers(n_workers)
clientend_date = parser.parse('2018-06-01')
start_date = parser.parse('2018-05-01')
nyc_tlc = NycTlcYellow(start_date=start_date, end_date=end_date)
nyc_tlc_df = nyc_tlc.to_pandas_dataframe()nyc_tlc_df.head()# As mentioned before, our VMs each have 2 GPUs, so we will partition among n_workers*2
df = dask_cudf.from_cudf(cudf.from_pandas(nyc_tlc_df), npartitions=n_workers * 2)# Query the dataframe to clean up the outliers
df = clean(df, must_haves)
# Add new features
taxi_df = df.map_partitions(add_features)
taxi_df = taxi_df.dropna()
taxi_df = taxi_df.astype("float32")
# Split into training and validation sets
X, y = taxi_df.drop(["fareAmount"], axis=1), taxi_df["fareAmount"].astype('float32')
X_train, X_test, y_train, y_test = train_test_split(X, y, shuffle=True)workers = client.has_what().keys()
X_train, y_train = dask_utils.persist_across_workers(client, [X_train, y_train], workers=workers)
cu_dask_rf = RandomForestRegressor(ignore_empty_partitions=True)
cu_dask_rf = cu_dask_rf.fit(X_train, y_train)
wait(cu_dask_rf.rfs)y_pred = cu_dask_rf.predict(X_test)
score = mean_squared_error(y_pred.compute().to_array(), y_test.compute().to_array())
print("Workflow Complete - RMSE: ", np.sqrt(score))client.close()
cluster.close() | 0 |
rapidsai_public_repos/cloud-ml-examples/azure/notebooks | rapidsai_public_repos/cloud-ml-examples/azure/notebooks/remote-explanation/azure-gpu-shap.ipynb | # %%bash
# apt-get update && \
# apt-get install -y fuse && \
# apt-get install -y build-essential && \
# apt-get install -y python3-dev && \
# pip install azureml-core && \
# pip install azureml-interpret && \
# pip install -e git+https://github.com/interpretml/interpret-community.git#egg=interpret_community\&subdirectory=python && \
# pip install raiwidgets# Check core SDK version number
import azureml.core
print("SDK version:", azureml.core.VERSION)# # Uncomment if you're using second method
# import os
# subscription_id = os.getenv("SUBSCRIPTION_ID", default="<subscription_ID>")
# resource_group = os.getenv("RESOURCE_GROUP", default="RAPIDS-SHAP")
# workspace_name = os.getenv("WORKSPACE_NAME", default="azure-intepret")
# workspace_region = os.getenv("WORKSPACE_REGION", default="eastus")# # Uncomment if you're using second method
# from azureml.core import Workspace
# try:
# ws = Workspace(subscription_id = subscription_id, resource_group = resource_group, workspace_name = workspace_name)
# # write the details of the workspace to a configuration file to the notebook library
# ws.write_config()
# print("Workspace configuration succeeded.")
# except:
# print("Workspace not accessible. Creting new workspace...")
# from azureml.core import Workspace
# # Create the workspace using the specified parameters
# ws = Workspace.create(name = workspace_name,
# subscription_id = subscription_id,
# resource_group = resource_group,
# location = workspace_region,
# create_resource_group = True,
# exist_ok = True)
# ws.get_details()
# # write the details of the workspace to a configuration file to the notebook library
# ws.write_config()from azureml.core import Workspace
ws = Workspace.from_config()
print(ws.name, ws.resource_group, ws.location, ws.subscription_id, sep='\n')
datastore = ws.get_default_datastore()
print("Default datastore's name: {}".format(datastore.name))from azureml.core import Experiment
experiment_name = 'gpu-shap-on-amlcompute'
experiment = Experiment(workspace=ws, name=experiment_name)from azureml.core.compute import ComputeTarget, AmlCompute
from azureml.core.compute_target import ComputeTargetException
# choose a name for your cluster
gpu_cluster_name = 'gpu-cluster'
if gpu_cluster_name in ws.compute_targets:
gpu_cluster = ws.compute_targets[gpu_cluster_name]
if gpu_cluster and type(gpu_cluster) is AmlCompute:
print('Found compute target. Will use {0} '.format(gpu_cluster_name))
else:
print('creating new cluster')
# m_size parameter below could be modified to one of the RAPIDS-supported VM types
provisioning_config = AmlCompute.provisioning_configuration(vm_size = 'Standard_NC6s_v3',
max_nodes = 1,
idle_seconds_before_scaledown = 300,
vm_priority = "lowpriority")
# Use VM types with more than one GPU for multi-GPU option, e.g. Standard_NC12s_v3
# create the cluster
gpu_cluster = ComputeTarget.create(ws, gpu_cluster_name, provisioning_config)
# can poll for a minimum number of nodes and for a specific timeout
# if no min node count is provided it uses the scale settings for the cluster
gpu_cluster.wait_for_completion(show_output=True, min_node_count=None, timeout_in_minutes=20)
# use get_status() to get a detailed status for the current cluster
print(gpu_cluster.get_status().serialize())from azureml.core import Environment
environment_name = "rapids"
env = Environment(environment_name)
env.docker.enabled = True
env.docker.base_image = None
#Installing interpret-community from source for now, will update later
env.docker.base_dockerfile = """
FROM rapidsai/rapidsai:21.06-cuda11.0-runtime-ubuntu18.04-py3.8
RUN apt-get update && \
apt-get install -y fuse && \
apt-get install -y build-essential && \
apt-get install -y python3-dev && \
source activate rapids && \
pip install azureml-defaults && \
pip install azureml-interpret && \
pip install -e git+https://github.com/interpretml/interpret-community.git#egg=interpret_community\&subdirectory=python && \
pip install azureml-telemetry
"""
env.python.user_managed_dependencies = Trueimport os
import shutil
project_folder = './scripts'
os.makedirs(project_folder, exist_ok=True)
shutil.copy('train_explain.py', project_folder)from azureml.core import Run
from azureml.core import ScriptRunConfig
src = ScriptRunConfig(source_directory=project_folder,
script='train_explain.py',
compute_target=gpu_cluster,
environment=env)
run = experiment.submit(config=src)
run%%time
# Shows output of the run on stdout.
run.wait_for_completion(show_output=True)run.get_metrics()from azureml.interpret import ExplanationClient
# Get model explanation data
client = ExplanationClient.from_run(run)
global_explanation = client.download_model_explanation()
local_importance_values = global_explanation.local_importance_values
expected_values = global_explanation.expected_values
# Or you can use the saved run.id to retrive the feature importance values
client = ExplanationClient.from_run_id(ws, experiment_name, run.id)
global_explanation = client.download_model_explanation()
local_importance_values = global_explanation.local_importance_values
expected_values = global_explanation.expected_values# Get the top k (e.g., 4) most important features with their importance values
global_explanation_topk = client.download_model_explanation(top_k=4)
global_importance_values = global_explanation_topk.get_ranked_global_values()
global_importance_names = global_explanation_topk.get_ranked_global_names()print('global importance values: {}'.format(global_importance_values))
print('global importance names: {}'.format(global_importance_names))# Retrieve model for visualization and deployment
from azureml.core.model import Model
import joblib
original_model = Model(ws, 'model_explain_model_on_amlcomp')
model_path = original_model.download(exist_ok=True)
original_model = joblib.load(model_path)# Retrieve x_test for visualization
import joblib
x_test_path = './x_test.pkl'
run.download_file('x_test_higgs.pkl', output_file_path=x_test_path)x_test = joblib.load('x_test.pkl')from interpret_community.widget import ExplanationDashboardimport cupy as cp
ExplanationDashboard(global_explanation, original_model,
datasetX=cp.asnumpy(x_test.values[:50])) | 0 |
rapidsai_public_repos/cloud-ml-examples/azure/notebooks | rapidsai_public_repos/cloud-ml-examples/azure/notebooks/remote-explanation/train_explain.py | #
# Copyright (c) 2021, NVIDIA CORPORATION.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
from interpret.ext.blackbox import TabularExplainer
from azureml.interpret import ExplanationClient
from cuml.model_selection import train_test_split
from azureml.core.run import Run
import joblib
import os
import cuml
from cuml.benchmark.datagen import load_higgs
OUTPUT_DIR = './outputs/'
os.makedirs(OUTPUT_DIR, exist_ok=True)
X, y = load_higgs()
N_ROWS = 1000000
run = Run.get_context()
client = ExplanationClient.from_run(run)
run.log('N_ROWS', N_ROWS)
X_train, X_test, y_train, y_test = train_test_split(X[:N_ROWS],
y[:N_ROWS],
random_state=1)
# write x_test out as a pickle file for later visualization
x_test_pkl = 'x_test.pkl'
with open(x_test_pkl, 'wb') as file:
joblib.dump(value=X_test, filename=os.path.join(OUTPUT_DIR, x_test_pkl))
run.upload_file('x_test_higgs.pkl', os.path.join(OUTPUT_DIR, x_test_pkl))
gamma = 0.001
C = 100.
# Use SVC algorithm to create a model
reg = cuml.svm.SVC(C=C, gamma=gamma, probability=True)
model = reg.fit(X_train, y_train)
# preds = reg.predict(X_test)
run.log('C', C)
run.log('gamma', gamma)
model_file_name = 'svc.pkl'
# save model in the outputs folder so it automatically get uploaded
with open(model_file_name, 'wb') as file:
joblib.dump(value=reg, filename=os.path.join(OUTPUT_DIR,
model_file_name))
# register the model
run.upload_file('original_model.pkl', os.path.join('./outputs/', model_file_name))
original_model = run.register_model(model_name='model_explain_model_on_amlcomp',
model_path='original_model.pkl')
# Explain predictions on your local machine
tabular_explainer = TabularExplainer(model, X_train.to_pandas(), features=X_train.columns, use_gpu=True)
# Explain overall model predictions (global explanation)
# Passing in test dataset for evaluation examples - note it must be a representative sample of the original data
# x_train can be passed as well, but with more examples explanations it will
# take longer although they may be more accurate
global_explanation = tabular_explainer.explain_global(X_test.to_pandas()[:50])
# Uploading model explanation data for storage or visualization in webUX
# The explanation can then be downloaded on any compute
comment = 'Global explanation on regression model trained on boston dataset'
client.upload_model_explanation(global_explanation, comment=comment, model_id=original_model.id)
| 0 |
rapidsai_public_repos/cloud-ml-examples/azure/notebooks | rapidsai_public_repos/cloud-ml-examples/azure/notebooks/configs/cloud_init.yaml.j2 | #cloud-config
# Bootstrap
packages:
- apt-transport-https
- ca-certificates
- curl
- gnupg-agent
- software-properties-common
- ubuntu-drivers-common
# Enable ipv4 forwarding, required on CIS hardened machines
write_files:
- path: /etc/sysctl.d/enabled_ipv4_forwarding.conf
content: |
net.ipv4.conf.all.forwarding=1
# create the docker group
groups:
- docker
# Add default auto created user to docker group
system_info:
default_user:
groups: [docker]
runcmd:
# Install Docker
- curl -fsSL https://download.docker.com/linux/ubuntu/gpg | apt-key add -
- add-apt-repository "deb [arch=amd64] https://download.docker.com/linux/ubuntu $(lsb_release -cs) stable"
- apt-get update -y
- apt-get install -y docker-ce docker-ce-cli containerd.io
- systemctl start docker
- systemctl enable docker
# Install NVIDIA driver
- DEBIAN_FRONTEND=noninteractive ubuntu-drivers install
# Install NVIDIA docker
- curl -fsSL https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
- curl -s -L https://nvidia.github.io/nvidia-docker/$(. /etc/os-release;echo $ID$VERSION_ID)/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
- apt-get update -y
- apt-get install -y nvidia-docker2
- systemctl restart docker
# Attempt to run a RAPIDS container to download the container layers and decompress them
- 'docker run --net=host --gpus=all --shm-size=256m rapidsai/rapidsai:cuda11.2-runtime-ubuntu18.04-py3.8 dask-scheduler --version'
| 0 |
rapidsai_public_repos/cloud-ml-examples/azure | rapidsai_public_repos/cloud-ml-examples/azure/kubernetes/Detailed_setup_guide.md | # [Detailed Guide to use Dask on Azure Kubernetes Service (AKS)](#anchor-start)
For all the next steps, we will be using the [Azure CLI](https://docs.microsoft.com/en-us/cli/azure/install-azure-cli), however the same can be achieved through the [Azure Portal](https://portal.azure.com/#home).
### [Step 0: Install and authenticate with Azure CLI](#anchor-install-azurecli)
- Install the `az` cli using
```
curl -sL https://aka.ms/InstallAzureCLIDeb | sudo bash
````
on the computer from where you will be running these examples from. You can remove the `sudo` if running inside a Docker container.
- Once `az` is installed, make sure you configure the local `az` cli to work with your Azure credentials, run `az login` and authenticate from Microsoft's website.
- For more details follow the steps [Here](https://docs.microsoft.com/en-us/cli/azure/install-azure-cli).
### [Step 1: Install Kubectl](#anchor-install-kubectl)
- Install `kubectl` to access your cluster from the local machine from the following link : [https://kubernetes.io/docs/tasks/tools/] depending on your operating system.
### [Step 2: Set some environment variables](#anchor-set-env-variables)
We will set some environment variables beforehand, namely to help us deploy some resources quickly. We will continue using common values for the rest of the deployments.
```bash
REGION_NAME=<your preferred location>
RESOURCE_GROUP=<your preferred resource group name>
SUBNET_NAME=<optional subnet name>
VNET_NAME=<optional vnet name>
VERSION=$(az aks get-versions --location $REGION_NAME --query \
'orchestrators[?!isPreview] | [-1].orchestratorVersion' \--output tsv)
AKS_CLUSTER_NAME=<your cluster name>
VM_SIZE=Standard_NC12s_v3 # or any other VM size. We use VMs with GPU
```
- We would need a resource group. You can create a resource group in Azure `az group create --name $RESOURCE_GROUP --location $REGION_NAME`, or use an existing one.
- Secondly, we get the latest non preview Kubernetes version for the specific region and store it in an env variable `VERSION`. At the time of writing this article, the latest version is 1.20.5. Optionally you can directly set the Kubernetes version in the environment variable `VERSION`.
- **NOTE 1**: There are two network modes to choose from when deploying an AKS cluster. The default one is *Kubenet networking* which we will use here.
- **NOTE 2**: Depending on your account limitations, the number and type of VMs that you can spin up may vary. Also there may be zone limitations. Make sure you spin up VMs with GPUs: NVIDIA Pascal™ or better with compute capability 6.0+. To give some examples of types of VMs you can use, the Azure [NC series](https://docs.microsoft.com/en-us/azure/virtual-machines/nc-series) / [NCv3 series](https://docs.microsoft.com/en-us/azure/virtual-machines/ncv3-series) VMs provide single or multi-gpu capabilities. In this setup guide for Kubernetes, we are using `Standard_NC12s_v3` VMs which have 2 NVIDIA V100 GPUs each.
### [Step 3: Create the cluster and get Kubernetes credentials](#anchor-create-aks-cluster)
Once you verify that you are allowed to use the necessary VM sizes in your preferred location, now its time to create a managed kubernetes cluster, namely an AKS cluster. The process is pretty simple. Also, after you successfully deploy a cluster with a node-pool of some nodes, you will be able to run workers as pods on the kubernetes cluster using [dask-kubernetes](https://github.com/dask/dask-kubernetes).
- Let's first run the following command to create a AKS cluster using the latest kubernetes version. It will take a few minutes before it completes. Grab a coffee :coffee: :coffee: .
```bash
az aks create \
--resource-group $RESOURCE_GROUP \
--name $AKS_CLUSTER_NAME \
--node-count 2 \
--location $REGION_NAME \
--kubernetes-version $VERSION \
--node-vm-size $VM_SIZE \
--generate-ssh-keys
```
- Once the cluster is created successfully, let's get the credentials for your AKS cluster to access it from your machine.
```bash
az aks get-credentials --resource-group $RESOURCE_GROUP --name $AKS_CLUSTER_NAME
```
- Check whether you are able to access the nodes:
```bash
kubectl get nodes
NAME STATUS ROLES AGE VERSION
aks-nodepool1-98672075-vmss000000 Ready agent 4m56s v1.20.5
aks-nodepool1-98672075-vmss000001 Ready agent 4m12s v1.20.5
```
### [Step 3: Set up the AKS cluster to use GPUs for our workload](#anchor-setup-gpu)
Once you have an AKS cluster up and running with nodes which have GPU capabilities, you need to install the [NVIDIA device plugin](https://github.com/NVIDIA/k8s-device-plugin) which allows allocation of GPUs to pods.
- First create a namespace using:
```
kubectl create namespace gpu-resources
```
- Create a file named *nvidia-device-plugin-ds.yaml* and paste the following manifest. This instruction set is taken from [Microsoft's official instructions](https://docs.microsoft.com/en-us/azure/aks/gpu-cluster). You can follow that as well.
```yaml
apiVersion: apps/v1
kind: DaemonSet
metadata:
name: nvidia-device-plugin-daemonset
namespace: gpu-resources
spec:
selector:
matchLabels:
name: nvidia-device-plugin-ds
updateStrategy:
type: RollingUpdate
template:
metadata:
# Mark this pod as a critical add-on; when enabled, the critical add-on scheduler
# reserves resources for critical add-on pods so that they can be rescheduled after
# a failure. This annotation works in tandem with the toleration below.
annotations:
scheduler.alpha.kubernetes.io/critical-pod: ""
labels:
name: nvidia-device-plugin-ds
spec:
tolerations:
# Allow this pod to be rescheduled while the node is in "critical add-ons only" mode.
# This, along with the annotation above marks this pod as a critical add-on.
- key: CriticalAddonsOnly
operator: Exists
- key: nvidia.com/gpu
operator: Exists
effect: NoSchedule
containers:
- image: mcr.microsoft.com/oss/nvidia/k8s-device-plugin:1.11
name: nvidia-device-plugin-ctr
securityContext:
allowPrivilegeEscalation: false
capabilities:
drop: ["ALL"]
volumeMounts:
- name: device-plugin
mountPath: /var/lib/kubelet/device-plugins
volumes:
- name: device-plugin
hostPath:
path: /var/lib/kubelet/device-plugins
```
Finally apply the NVIDIA Device plugin so that the pods can see the GPU :
```bash
kubectl apply -f nvidia-device-plugin-ds.yml
```
### [Step 4: Create Azure Container Registry for pulling and pushing worker and scheduler docker images](#anchor-setup-azure-container-repository)
We will also need a container registry for the pod images. Here, we will use the container repository provided by Azure (ACR).
- Create an Azure container repository env variable which will be useful later.
```bash
ACR_NAME=<your repo name>
```
- Create an Azure Container Registry in the same region and under the same resource group with the Standard SKU with the following:
```bash
az acr create \
--resource-group $RESOURCE_GROUP \
--location $REGION_NAME \
--name $ACR_NAME \
--sku Standard
```
### [Step 5: Authenticate AKS to pull images from ACR using secrets](#anchor-setup-aks-acr-authentication)
We need to authenticate AKS to pull the images from ACR. For this purpose we will pass a secret in the pod configurations. We need to perform the following steps for that
- Admin enable ACR
```
az acr update -n $ACR_NAME --admin-enabled true
```
- Get the username and password using of ACR using
```
>> az acr credential show --name $ACR_NAME
{
"passwords": [
{
"name": "password",
"value": "<some password 2>"
},
{
"name": "password2",
"value": "<some password 2>"
}
],
"username": "<your user name>"
}
```
Note down the password and the username. Any of the two passwords will work.
- Docker login and create a secret with `docker login`
```bash
docker login $ACR_NAME.azurecr.io
kubectl create secret docker-registry aks-secret --docker-server=$ACR_NAME.azurecr.io \
--docker-username=$ACR_NAME --docker-password=<passwd> --docker-email=<any-email>
```
- **IMPORTANT:** The password/credentials for ACR expires every 3 hours. Make sure you renew them from the Azure portal if you want AKS to pull the images during pod creation. Then update the docker secret accordingly before scaling up or starting the cluster.
- And then pass the secret to a pod manifest (all pod specification `yaml` files are in `./podspecs` directory, replace necessary details in those files) like the following:
```yaml
kind: Pod
spec:
restartPolicy: Never
containers:
- image: <user>.azurecr.io/<image-path>
imagePullPolicy: IfNotPresent
name: dask-scheduler
resources:
limits:
cpu: "4"
memory: 25G
requests:
cpu: "4"
memory: 25G
imagePullSecrets:
- name: "aks-secret"
```
### [Step 6: Build and push the pod images to ACR](#anchor-build-podimages)
Now we build and push the pod images to ACR. We have a `Dockerfile` in the current directory. From the current directory do the following:
```bash
docker build -t $ACR_NAME.azurecr.io/aks-mnmg/dask-unified:21.06 .
docker push $ACR_NAME.azurecr.io/aks-mnmg/dask-unified:21.06
```
### [Step 7: Install dask-kubernetes python library if not already present](#anchor-install-daskcloudprovider)
Install [dask-kubernetes](https://kubernetes.dask.org/en/latest/) if not already installed.
| 0 |
Subsets and Splits