metadata
dataset_info:
features:
- name: id
dtype: int64
- name: alt_text
dtype: string
- name: header
dtype: string
- name: transcript
dtype: string
splits:
- name: train
num_bytes: 3789735
num_examples: 3179
- name: test
num_bytes: 200108
num_examples: 168
download_size: 2546922
dataset_size: 3989843
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*