README.md

---
dataset_info:
  features:
  - name: license
    dtype: int64
  - name: file_name
    dtype: string
  - name: coco_url
    dtype: string
  - name: height
    dtype: int64
  - name: width
    dtype: int64
  - name: date_captured
    dtype: string
  - name: flickr_url
    dtype: string
  - name: id
    dtype: int64
  splits:
  - name: train
    num_bytes: 5016546
    num_examples: 25000
  download_size: 1741047
  dataset_size: 5016546
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---