metadata
dataset_info:
features:
- name: text
dtype: string
- name: original_captions
sequence: string
- name: category
dtype: int64
splits:
- name: train
num_bytes: 4780484
num_examples: 4217
download_size: 2368379
dataset_size: 4780484
configs:
- config_name: default
data_files:
- split: train
path: data/train-*