File size: 2,913 Bytes
9a997e4 a241bb3 b47829b a241bb3 b47829b 8e0d56d 18ba8c1 8e0d56d a241bb3 8e0d56d a241bb3 9a997e4 615cfe4 a241bb3 9a997e4 b47829b 74c0c8e 9a997e4 8e0d56d b47829b 8e0d56d 9a997e4 74c0c8e 9a997e4 18ba8c1 9a997e4 18ba8c1 9a997e4 8e0d56d a241bb3 8e0d56d 9a997e4 8e0d56d 9a997e4 8e0d56d 18ba8c1 8e0d56d 9a997e4 8e0d56d 9a997e4 b47829b 9a997e4 b47829b 9a997e4 b47829b 9a997e4 b47829b 9a997e4 b0303a0 9a997e4 b47829b 9a997e4 ec21179 b47829b 9a997e4 18ba8c1 8e0d56d 18ba8c1 8e0d56d 18ba8c1 8e0d56d 18ba8c1 8e0d56d 9a997e4 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 |
"""Train and compile the model."""
import shutil
import numpy
import pandas
import pickle
from settings import (
DEPLOYMENT_PATH,
DATA_PATH,
INPUT_SLICES,
PRE_PROCESSOR_APPLICANT_PATH,
PRE_PROCESSOR_BANK_PATH,
PRE_PROCESSOR_CREDIT_BUREAU_PATH,
APPLICANT_COLUMNS,
BANK_COLUMNS,
CREDIT_BUREAU_COLUMNS,
)
from utils.client_server_interface import MultiInputsFHEModelDev
from utils.model import MultiInputDecisionTreeClassifier
from utils.pre_processing import get_pre_processors
def get_multi_inputs(data):
"""Get inputs for all three parties from the input data, using fixed slices.
Args:
data (numpy.ndarray): The input data to consider.
Returns:
(Tuple[numpy.ndarray]): The inputs for all three parties.
"""
return (
data[:, INPUT_SLICES["applicant"]],
data[:, INPUT_SLICES["bank"]],
data[:, INPUT_SLICES["credit_bureau"]]
)
print("Load and pre-process the data")
# Load the data
data = pandas.read_csv(DATA_PATH, encoding="utf-8")
# Define input and target data
data_x = data.copy()
data_y = data_x.pop("Target").copy().to_frame()
# Get data from all parties
data_applicant = data_x[APPLICANT_COLUMNS].copy()
data_bank = data_x[BANK_COLUMNS].copy()
data_credit_bureau = data_x[CREDIT_BUREAU_COLUMNS].copy()
# Feature engineer the data
pre_processor_applicant, pre_processor_bank, pre_processor_credit_bureau = get_pre_processors()
preprocessed_data_applicant = pre_processor_applicant.fit_transform(data_applicant)
preprocessed_data_bank = pre_processor_bank.fit_transform(data_bank)
preprocessed_data_credit_bureau = pre_processor_credit_bureau.fit_transform(data_credit_bureau)
preprocessed_data_x = numpy.concatenate((preprocessed_data_applicant, preprocessed_data_bank, preprocessed_data_credit_bureau), axis=1)
print("\nTrain and compile the model")
model = MultiInputDecisionTreeClassifier()
model, sklearn_model = model.fit_benchmark(preprocessed_data_x, data_y)
multi_inputs_train = get_multi_inputs(preprocessed_data_x)
model.compile(*multi_inputs_train, inputs_encryption_status=["encrypted", "encrypted", "encrypted"])
print("\nSave deployment files")
# Delete the deployment folder and its content if it already exists
if DEPLOYMENT_PATH.is_dir():
shutil.rmtree(DEPLOYMENT_PATH)
# Save files needed for deployment (and enable cross-platform deployment)
fhe_model_dev = MultiInputsFHEModelDev(DEPLOYMENT_PATH, model)
fhe_model_dev.save(via_mlir=True)
# Save pre-processors
with (
PRE_PROCESSOR_APPLICANT_PATH.open('wb') as file_applicant,
PRE_PROCESSOR_BANK_PATH.open('wb') as file_bank,
PRE_PROCESSOR_CREDIT_BUREAU_PATH.open('wb') as file_credit_bureau,
):
pickle.dump(pre_processor_applicant, file_applicant)
pickle.dump(pre_processor_bank, file_bank)
pickle.dump(pre_processor_credit_bureau, file_credit_bureau)
print("\nDone !")
|