metadata
dataset_info:
features:
- name: image
dtype: image
- name: image_id
dtype: int64
- name: question
dtype: string
- name: answer
dtype: string
splits:
- name: train
num_bytes: 1274978820.86
num_examples: 5742
download_size: 219342715
dataset_size: 1274978820.86
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
The dataset now includes a sample set of 5.7K entries from the full dataset of 85K, which will be made publicly available once finalized!