metadata
dataset_info:
features:
- name: id
dtype: string
- name: images
list:
- name: bytes
dtype: binary
- name: path
dtype: string
- name: conversations
list:
- name: content
dtype: string
- name: role
dtype: string
- name: source
dtype: string
splits:
- name: train
num_bytes: 3064002869
num_examples: 1000
download_size: 3063812668
dataset_size: 3064002869
configs:
- config_name: default
data_files:
- split: train
path: data/train-*