metadata
dataset_info:
features:
- name: id
dtype: string
- name: images
list:
- name: bytes
dtype: 'null'
- name: path
dtype: string
- name: conversation
list:
- name: content
dtype: string
- name: role
dtype: string
splits:
- name: train
num_bytes: 32321172
num_examples: 23201
- name: eval
num_bytes: 8707158
num_examples: 6249
- name: test
num_bytes: 19116592
num_examples: 15480
download_size: 2724018
dataset_size: 60144922
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: eval
path: data/eval-*
- split: test
path: data/test-*