metadata
dataset_info:
features:
- name: id
dtype: int32
- name: image
dtype: image
- name: caption
dtype: string
- name: caption_ls
list: string
- name: category
dtype: string
- name: filepath
dtype: string
- name: sentids
list: int32
- name: filename
dtype: string
- name: split
dtype: string
- name: sentences_tokens
list:
list: string
- name: en_sentences_raw
list: string
- name: sentences_sentid
list: int32
- name: cocoid
dtype: int32
splits:
- name: train
num_bytes: 18640275938.125
num_examples: 113287
- name: validation
num_bytes: 819179714
num_examples: 5000
- name: test
num_bytes: 826633941
num_examples: 5000
download_size: 20164806114
dataset_size: 20286089593.125
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*