---
dataset_info:
  features:
  - name: audio
    dtype:
      audio:
        sampling_rate: 16000
  - name: transkription
    dtype: string
  - name: source
    dtype: string
  - name: predicted_transkription
    dtype: string
  splits:
  - name: train
    num_bytes: 20603875516.34
    num_examples: 546561
  download_size: 21827534712
  dataset_size: 20603875516.34
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---