metadata
dataset_info:
features:
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
- name: images
list: image
splits:
- name: train
num_bytes: 6781367
num_examples: 16
download_size: 6764781
dataset_size: 6781367
configs:
- config_name: default
data_files:
- split: train
path: data/train-*