metadata
dataset_info:
features:
- name: image
dtype: image
- name: url
dtype: string
- name: caption
dtype: string
- name: id
dtype: int64
- name: similarity
dtype: float64
splits:
- name: train
num_bytes: 343403671.807
num_examples: 5117
download_size: 282913216
dataset_size: 343403671.807
configs:
- config_name: default
data_files:
- split: train
path: data/train-*