README.md

metadata

dataset_info:
  features:
    - name: arxiv_id
      dtype: string
    - name: tables
      list:
        - name: index
          dtype: int64
        - name: records
          list:
            - name: column
              dtype: int64
            - name: dataset
              dtype: string
            - name: metric
              dtype: string
            - name: model
              dtype: string
            - name: row
              dtype: int64
            - name: task
              dtype: string
            - name: value
              dtype: string
  splits:
    - name: train
      num_bytes: 539137
      num_examples: 731
  download_size: 108254
  dataset_size: 539137
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

This dataset was proposed in AxCell: Automatic Extraction of Results from Machine Learning Papers.