README.md

metadata

license: apache-2.0
size_categories:
  - 1K<n<10K
task_categories:
  - object-detection
pretty_name: Aerial Image Data
annotations: images/metadata.jsonl
tags:
  - climate
dataset_info:
  - config_name: aug_full_features
    features:
      - name: image
        dtype: string
      - name: image_id
        dtype: string
      - name: width
        dtype: int64
      - name: height
        dtype: int64
      - name: good
        dtype: bool
      - name: blank
        dtype: bool
      - name: objects
        struct:
          - name: area
            sequence: float64
          - name: bbox
            sequence:
              sequence: float64
          - name: category
            sequence: int64
          - name: id
            sequence: int64
    splits:
      - name: train
        num_bytes: 3539261
        num_examples: 10249
    download_size: 1034461
    dataset_size: 3539261
  - config_name: full_features
    features:
      - name: image
        dtype: string
      - name: image_id
        dtype: string
      - name: width
        dtype: int64
      - name: height
        dtype: int64
      - name: good
        dtype: bool
      - name: blank
        dtype: bool
      - name: objects
        struct:
          - name: area
            sequence: int64
          - name: bbox
            sequence:
              sequence: float64
          - name: category
            sequence: int64
          - name: id
            sequence: int64
    splits:
      - name: train
        num_bytes: 1790034
        num_examples: 4772
    download_size: 487822
    dataset_size: 1790034
  - config_name: full_features_fake
    features:
      - name: image
        dtype: string
      - name: image_id
        dtype: string
      - name: width
        dtype: int64
      - name: height
        dtype: int64
      - name: good
        dtype: bool
      - name: blank
        dtype: bool
      - name: objects
        struct:
          - name: area
            sequence: int64
          - name: bbox
            sequence:
              sequence: float64
          - name: category
            sequence: int64
          - name: id
            sequence: int64
    splits:
      - name: train
        num_bytes: 1928454
        num_examples: 4772
    download_size: 484928
    dataset_size: 1928454
  - config_name: metadata
    features:
      - name: file_name
        dtype: string
      - name: bbox
        dtype: string
      - name: categories
        dtype: int64
      - name: id
        dtype: int64
      - name: good
        dtype: bool
      - name: width
        dtype: int64
      - name: height
        dtype: int64
      - name: area
        dtype: int64
      - name: category_name
        dtype: string
      - name: image_id
        dtype: string
      - name: blank
        dtype: bool
    splits:
      - name: train
        num_bytes: 3102587
        num_examples: 26156
    download_size: 1285477
    dataset_size: 3102587
configs:
  - config_name: aug_full_features
    data_files:
      - split: train
        path: aug_full_features/train-*
  - config_name: full_features
    data_files:
      - split: train
        path: full_features/train-*
  - config_name: full_features_fake
    data_files:
      - split: train
        path: full_features_fake/train-*
  - config_name: metadata
    data_files:
      - split: train
        path: metadata/train-*