metadata
dataset_info:
features:
- name: filename
dtype: string
- name: captions
sequence: string
- name: image
dtype: image
splits:
- name: train
num_bytes: 454898418.085
num_examples: 2695
- name: test
num_bytes: 40259984
num_examples: 231
- name: valid
num_bytes: 161459658
num_examples: 924
download_size: 663096627
dataset_size: 656618060.085
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*