metadata

dataset_info:
  features:
    - name: audio
      dtype:
        audio:
          sampling_rate: 16000
    - name: transkription
      dtype: string
    - name: source
      dtype: string
    - name: predicted_transkription
      dtype: string
    - name: cer
      dtype: float64
  splits:
    - name: train_tuda_0
      num_bytes: 13757929586.4
      num_examples: 42812
    - name: train_cv19_0
      num_bytes: 20608244029.34
      num_examples: 546561
    - name: train_emilia_0
      num_bytes: 34405160493
      num_examples: 627354
    - name: train_mls_0
      num_bytes: 102822624468.044
      num_examples: 390578
    - name: train_emilia_yodas0
      num_bytes: 158522556378
      num_examples: 1980468
    - name: train_eurospeech
      num_bytes: 243362644825.86
      num_examples: 502515
    - name: train_el_tts
      num_bytes: 123890289
      num_examples: 495
  download_size: 570166922667
  dataset_size: 573603050069.644
configs:
  - config_name: default
    data_files:
      - split: train_tuda_0
        path: data/train_tuda_0-*
      - split: train_cv19_0
        path: data/train_cv19_0-*
      - split: train_emilia_0
        path: data/train_emilia_0-*
      - split: train_mls_0
        path: data/train_mls_0-*
      - split: train_emilia_yodas0
        path: data/train_emilia_yodas0-*
      - split: train_el_tts
        path: data/train_el_tts-*
      - split: train_eurospeech
        path: data/train_eurospeech-*
license: cc-by-nc-4.0
task_categories:
  - automatic-speech-recognition
language:
  - de
pretty_name: German whisper labeled ASR Collection

Dataset Card

Dataset Sources and Licensing

This dataset is a mixture of several German and multilingual speech datasets. For each dataset, the license of the original author applies. Please consult the linked sources for detailed licensing information and terms of use.

Dataset Name	Original Source / Author	Link
TUDA-De German Speech Corpus	LT Group at UHH / TU Darmstadt	https://huggingface.co/datasets/uhhlt/Tuda-De
Mozilla Common Voice	Mozilla Foundation	https://commonvoice.mozilla.org/en/datasets
Emilia Speech Dataset	Amphion	https://huggingface.co/datasets/amphion/Emilia-Dataset
Emilia Yodas Speech Dataset	Amphion	https://huggingface.co/datasets/amphion/Emilia-Dataset
Multilingual LibriSpeech (MLS)	Facebook AI Research	https://arxiv.org/abs/2012.03411
EuroSpeech Dataset	ETH Zurich / disco-eth	https://huggingface.co/datasets/disco-eth/EuroSpeech

Note: If a dataset does not have a public source listed, please contact the dataset provider or refer to your data distributor for licensing details.

Licensing Disclaimer

The use of this dataset and any derived models must comply with the licenses of the original datasets. The dataset creators do not claim any rights to the datasets themselves. Always verify the license terms before using the data or models for commercial or research purposes.