metadata
dataset_info:
features:
- name: images
sequence: image
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
splits:
- name: train
num_bytes: 152248793455.6
num_examples: 950
- name: test
num_bytes: 8013094392.4
num_examples: 50
download_size: 164605519
dataset_size: 160261887848
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*