metadata
dataset_info:
features:
- name: image
dtype: image
- name: conversations
dtype: string
- name: texts
list:
- name: from
dtype: string
- name: value
dtype: string
splits:
- name: train
num_bytes: 20763027848
num_examples: 272795
download_size: 20388488209
dataset_size: 20763027848
configs:
- config_name: default
data_files:
- split: train
path: data/train-*