metadata
dataset_info:
features:
- name: dataset_id
dtype: string
- name: image_id
dtype: string
- name: annotator_id
dtype: int64
- name: caption
dtype: string
- name: segmented_caption
sequence: string
splits:
- name: valid
num_bytes: 15307808
num_examples: 41691
- name: test
num_bytes: 47335910
num_examples: 126020
download_size: 25740200
dataset_size: 62643718
configs:
- config_name: default
data_files:
- split: valid
path: data/valid-*
- split: test
path: data/test-*