metadata
dataset_info:
features:
- name: conversations
list:
- name: content
dtype: string
- name: role
dtype: string
- name: image
dtype: image
splits:
- name: train
num_bytes: 6553159165
num_examples: 595375
download_size: 6491476302
dataset_size: 6553159165
configs:
- config_name: default
data_files:
- split: train
path: data/train-*