metadata
dataset_info:
features:
- name: caption
dtype: string
- name: url
dtype: string
- name: translated_caption
dtype: string
- name: cosine_similarity
dtype: float64
splits:
- name: train
num_bytes: 1211184214
num_examples: 3552701
download_size: 750080206
dataset_size: 1211184214
configs:
- config_name: default
data_files:
- split: train
path: data/train-*