metadata
dataset_info:
features:
- name: answer
dtype: string
- name: category
dtype: string
- name: qid
dtype: string
- name: images
list: image
- name: problem
dtype: string
- name: id
dtype: int64
splits:
- name: train
num_bytes: 2938932145.25
num_examples: 38870
download_size: 2898144249
dataset_size: 2938932145.25
configs:
- config_name: default
data_files:
- split: train
path: data/train-*