---
dataset_info:
  features:
  - name: audio_tokens
    sequence: int64
  - name: raw_text
    dtype: string
  - name: transcript
    dtype: string
  - name: entities
    sequence: string
  - name: prompt
    dtype: string
  - name: input_ids
    sequence: int64
  - name: attention_mask
    sequence: int64
  splits:
  - name: train
    num_bytes: 2895627778
    num_examples: 70397
  download_size: 184378881
  dataset_size: 2895627778
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---