---
dataset_info:
  features:
  - name: input_ids
    list: int32
  - name: attention_mask
    list: int8
  - name: digit_labels
    list: int64
  - name: num_latents
    dtype: int64
  - name: K_bucket
    dtype: string
  splits:
  - name: train
    num_bytes: 26325175
    num_examples: 52582
  - name: eval
    num_bytes: 554110
    num_examples: 1000
  download_size: 6213152
  dataset_size: 26879285
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: eval
    path: data/eval-*
---