Datasets:

Modalities:
Image
Text
Formats:
parquet
Libraries:
Datasets
Dask
shunk031's picture
Upload dataset
7a0ca34 verified
metadata
dataset_info:
  - config_name: default
    features:
      - name: paper_id
        dtype: string
      - name: title
        dtype: string
      - name: abstract
        dtype: string
      - name: GA_figure
        dtype: image
      - name: GA_caption
        dtype: string
      - name: figures
        sequence:
          - name: image
            dtype: image
          - name: caption
            dtype: string
      - name: research_fields
        sequence: string
      - name: journal_ref
        dtype: string
      - name: conference
        dtype: string
    splits:
      - name: train
        num_bytes: 98581891115.832
        num_examples: 14331
      - name: validation
        num_bytes: 5390595408.112
        num_examples: 3096
      - name: test
        num_bytes: 5555685104.104
        num_examples: 3096
    download_size: 56137334585
    dataset_size: 109528171628.048
  - config_name: inpainting
    features:
      - name: paper_id
        dtype: string
      - name: GA_figure
        dtype: image
      - name: GA_figure_mask
        dtype: image
    splits:
      - name: train
        num_bytes: 4154125273.097
        num_examples: 14331
      - name: validation
        num_bytes: 830668643.76
        num_examples: 3096
      - name: test
        num_bytes: 844843415.568
        num_examples: 3096
    download_size: 5891524186
    dataset_size: 5829637332.425
  - config_name: ocr
    features:
      - name: paper_id
        dtype: string
      - name: GA_figure_ocr
        list:
          - name: text
            dtype: string
          - name: bounds
            struct:
              - name: bottomRight
                sequence: int64
              - name: topRight
                sequence: int64
              - name: bottomLeft
                sequence: int64
              - name: topLeft
                sequence: int64
          - name: langs
            list:
              - name: confidence
                dtype: float64
              - name: languageCode
                dtype: string
          - name: orientation
            dtype: string
          - name: confidence
            dtype: float64
      - name: figures
        sequence:
          - name: ocr
            list:
              - name: text
                dtype: string
              - name: bounds
                struct:
                  - name: bottomRight
                    sequence: int64
                  - name: topRight
                    sequence: int64
                  - name: bottomLeft
                    sequence: int64
                  - name: topLeft
                    sequence: int64
              - name: langs
                list:
                  - name: confidence
                    dtype: float64
                  - name: languageCode
                    dtype: string
              - name: orientation
                dtype: string
              - name: confidence
                dtype: float64
    splits:
      - name: train
        num_bytes: 383111069
        num_examples: 14331
      - name: validation
        num_bytes: 61814851
        num_examples: 3096
      - name: test
        num_bytes: 61518918
        num_examples: 3096
    download_size: 104386270
    dataset_size: 506444838
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*
  - config_name: inpainting
    data_files:
      - split: train
        path: inpainting/train-*
      - split: validation
        path: inpainting/validation-*
      - split: test
        path: inpainting/test-*
  - config_name: ocr
    data_files:
      - split: train
        path: ocr/train-*
      - split: validation
        path: ocr/validation-*
      - split: test
        path: ocr/test-*