metadata
dataset_info:
features:
- name: messages
dtype: string
- name: images
sequence: image
splits:
- name: train
num_bytes: 197196363
num_examples: 4000
download_size: 191694906
dataset_size: 197196363
configs:
- config_name: default
data_files:
- split: train
path: data/train-*