metadata
dataset_info:
features:
- name: id
dtype: int32
- name: image
dtype: image
- name: conversations
list:
- name: role
dtype: string
- name: content
dtype: string
splits:
- name: train
num_bytes: 489521323612.25
num_examples: 8689566
- name: validation
num_bytes: 6199537183.625
num_examples: 382403
download_size: 467383893861
dataset_size: 495720860795.875
configs:
- config_name: default
data_files:
- split: train
path: chat/train-*
- split: validation
path: chat/validation-*