pdf-ocr-dataset / README.md
broadfield-dev's picture
Upload dataset
1b4e7a0 verified
|
raw
history blame
703 Bytes
metadata
dataset_info:
  features:
    - name: documents
      dtype: string
    - name: embeddings
      sequence: float64
    - name: metadatas
      struct:
        - name: corrected_text
          dtype: string
        - name: corrections
          dtype: string
        - name: filename
          dtype: string
        - name: page_file
          dtype: string
        - name: page_num
          dtype: int64
        - name: words
          dtype: string
    - name: ids
      dtype: string
    - name: urls
      dtype: string
    - name: processed_at
      dtype: string
  splits:
    - name: train
      num_bytes: 18079
      num_examples: 1
  download_size: 37316
  dataset_size: 18079
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*