metadata
dataset_info:
features:
- name: caption
dtype: string
- name: caption_embs
sequence:
sequence: float16
- name: dataset_types
dtype: string
splits:
- name: train
num_bytes: 118658850
num_examples: 1000
download_size: 111700173
dataset_size: 118658850
configs:
- config_name: default
data_files:
- split: train
path: data/train-*