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: en_caption
list: string
- name: metadata
struct:
- name: description
dtype: string
- name: version
dtype: string
- name: data_year
dtype: string
- name: main_category
dtype: string
- name: width
dtype: int32
- name: height
dtype: int32
- name: file_name
dtype: string
- name: supercategory
dtype: string
splits:
- name: train
num_bytes: 1172596975755.25
num_examples: 2148214
- name: validation
num_bytes: 160798762960.625
num_examples: 267011
download_size: 1319632941418
dataset_size: 1333395738715.875
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*