tokenized-km-news / README.md
mlao01's picture
Upload dataset
d1828ad verified
metadata
dataset_info:
  features:
    - name: input_text
      dtype: string
    - name: input_ids
      sequence: int32
    - name: attention_mask
      sequence: int8
  splits:
    - name: train
      num_bytes: 8766686434
      num_examples: 2050000
    - name: validation
      num_bytes: 60369417
      num_examples: 14078
  download_size: 2251755758
  dataset_size: 8827055851
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*