metadata
dataset_info:
features:
- name: images
list: 'null'
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
- name: id
dtype: string
- name: source_dataset
dtype: string
splits:
- name: train
num_bytes: 26234500
num_examples: 17440
download_size: 8939232
dataset_size: 26234500
configs:
- config_name: default
data_files:
- split: train
path: data/train-*