metadata
dataset_info:
features:
- name: images
sequence: image
- name: texts
list:
- name: user
dtype: string
- name: assistant
dtype: string
- name: source
dtype: string
- name: image
dtype: image
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
splits:
- name: train
num_bytes: 36331050941.248
num_examples: 56398
download_size: 36000592447
dataset_size: 36331050941.248
configs:
- config_name: default
data_files:
- split: train
path: data/train-*