README.md

---
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: 6313382833354.764
    num_examples: 2386486
  - name: validation
    num_bytes: 769032107713.42
    num_examples: 297330
  download_size: 7069777984954
  dataset_size: 7082414941068.184
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
---