metadata
dataset_info:
features:
- name: caption
dtype: string
- name: caption_embs
sequence: float16
- name: dataset_types
dtype: string
splits:
- name: train
num_bytes: 46648211
num_examples: 42581
download_size: 43861275
dataset_size: 46648211
configs:
- config_name: default
data_files:
- split: train
path: data/train-*