metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: caption
dtype: string
- name: image_data
dtype: image
- name: image_embeddings
sequence: float32
- name: text_embeddings
sequence: float32
splits:
- name: train
num_bytes: 18422754083.25
num_examples: 143134
download_size: 18711756794
dataset_size: 18422754083.25