metadata
dataset_info:
features:
- name: orig_id
dtype: int64
- name: image
dtype: image
- name: caption
dtype: string
- name: label
dtype: int64
- name: image_idx
dtype: int64
- name: caption_idx
dtype: int64
splits:
- name: train
num_bytes: 2074327689
num_examples: 1200
- name: validation
num_bytes: 745577709
num_examples: 400
download_size: 2819418566
dataset_size: 2819905398
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*