tokenized-wikipedia / README.md
Q-bert's picture
Upload dataset
381472e
|
raw
history blame
446 Bytes
metadata
license: mit
dataset_info:
  features:
    - name: id
      dtype: int64
    - name: text
      dtype: string
    - name: title
      dtype: string
    - name: input_ids
      sequence: int32
    - name: attention_mask
      sequence: int8
  splits:
    - name: train
      num_bytes: 6518515310
      num_examples: 524601
  download_size: 1183778649
  dataset_size: 6518515310
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*