ytzi's picture
pre-tokenized with gpt2 (part 00002-of-00003)
fb90bc8 verified
metadata
dataset_info:
  features:
    - name: content
      dtype: string
    - name: input_ids
      sequence: int32
    - name: ratio_char_token
      dtype: float64
    - name: token_count
      dtype: int64
  splits:
    - name: train
      num_bytes: 63883229321
      num_examples: 12725978
  download_size: 20986659791
  dataset_size: 63883229321
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*