metadata
dataset_info:
features:
- name: id
dtype: string
- name: image_path
dtype: string
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
- name: image
dtype: image
splits:
- name: train
num_bytes: 11610356976
num_examples: 10000
download_size: 11594179084
dataset_size: 11610356976
configs:
- config_name: default
data_files:
- split: train
path: data/train-*