metadata
dataset_info:
features:
- name: image
dtype: image
- name: index
dtype: int64
- name: source_dataset
dtype: string
- name: caption
dtype: string
- name: question
dtype: string
- name: image_source
dtype: string
- name: question_id
dtype: string
- name: capability
dtype: string
- name: answer
dtype: string
- name: responses
dtype: string
splits:
- name: train
num_bytes: 255213080
num_examples: 198
download_size: 254965118
dataset_size: 255213080
configs:
- config_name: default
data_files:
- split: train
path: data/train-*