metadata
dataset_info:
features:
- name: image
dtype: image
- name: idx
dtype: int64
- name: word
dtype: string
splits:
- name: train
num_bytes: 206639400
num_examples: 100000
- name: test
num_bytes: 19683520
num_examples: 10000
download_size: 173346403
dataset_size: 226322920
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*