metadata
dataset_info:
features:
- name: text
dtype: string
- name: label
dtype: int64
- name: label_name
dtype: string
- name: caption
dtype: string
splits:
- name: train
num_bytes: 26099837
num_examples: 27250
download_size: 12114215
dataset_size: 26099837
configs:
- config_name: default
data_files:
- split: train
path: data/train-*