metadata
dataset_info:
features:
- name: total_text_length
dtype: int64
- name: total_img_length
dtype: int64
- name: image_size
dtype: int64
- name: text_lengths
sequence: int64
splits:
- name: train
num_bytes: 3509768
num_examples: 10000
download_size: 862090
dataset_size: 3509768
configs:
- config_name: default
data_files:
- split: train
path: data/train-*