metadata
dataset_info:
features:
- name: image
dtype: image
- name: conversation
sequence:
- name: role
dtype: string
- name: content
sequence:
- name: type
dtype: string
- name: text
dtype: string
splits:
- name: train
num_bytes: 13980051941.88
num_examples: 4320
- name: test
num_bytes: 1713304975
num_examples: 540
- name: val
num_bytes: 1717008668
num_examples: 540
download_size: 17342346116
dataset_size: 17410365584.879997
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: val
path: data/val-*