pretraining_data / README.md
qxuanson's picture
Upload dataset
ebdfb22 verified
|
raw
history blame
635 Bytes
metadata
dataset_info:
  features:
    - name: _id
      struct:
        - name: $oid
          dtype: string
    - name: url
      dtype: string
    - name: title
      dtype: string
    - name: content
      dtype: string
    - name: domain
      dtype: string
    - name: category
      sequence: string
    - name: create_at
      struct:
        - name: $date
          dtype: string
    - name: _class
      dtype: string
    - name: __index_level_0__
      dtype: int64
  splits:
    - name: train
      num_bytes: 86117186
      num_examples: 23637
  download_size: 28943868
  dataset_size: 86117186
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*