metadata
dataset_info:
features:
- name: image
dtype: image
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
splits:
- name: train
num_bytes: 2260324942.832
num_examples: 7168
- name: test
num_bytes: 286716500
num_examples: 891
- name: val
num_bytes: 263209858
num_examples: 859
download_size: 2796815792
dataset_size: 2810251300.832
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: val
path: data/val-*