---
dataset_info:
  features:
  - name: embed_text
    dtype: string
  - name: input_ids
    sequence: int32
  - name: token_type_ids
    sequence: int8
  - name: attention_mask
    sequence: int8
  - name: __index_level_0__
    dtype: int64
  splits:
  - name: train
    num_bytes: 13563865198
    num_examples: 5627729
  download_size: 4385109525
  dataset_size: 13563865198
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---