metadata
dataset_info:
features:
- name: images
list: image
- name: conversations
list:
- name: user
dtype: string
- name: assistant
dtype: string
- name: __index_level_0__
dtype: int64
splits:
- name: train
num_bytes: 2012107
num_examples: 50
download_size: 1931454
dataset_size: 2012107
configs:
- config_name: default
data_files:
- split: train
path: data/train-*