README.md

metadata

dataset_info:
  features:
    - name: text
      dtype: string
    - name: len_tokens
      dtype: int64
  splits:
    - name: train
      num_bytes: 44143916
      num_examples: 48016
  download_size: 8526144
  dataset_size: 44143916
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

#tokenizer: google/gemma-2b-it

hist len_tokens

describe.

percentil.