README.md

metadata

language:
  - da
license: cc0-1.0
tasks:
  - automatic-speech-recognition
datasets:
  - common_voice_8_0
metrics:
  - wer
model-index:
  - name: kblab-voxrex-wav2vec2-large-cv8-da
    results:
      - task:
          type: automatic-speech-recognition
        dataset:
          type: mozilla-foundation/common_voice_8_0
          args: da
          name: Danish Common Voice 8.0
        metrics:
          - type: wer
            value: 30.51
      - task:
          type: automatic-speech-recognition
        dataset:
          type: Alvenir/alvenir_asr_da_eval
          name: Alvenir ASR test dataset
        metrics:
          - type: wer
            value: 28.33

KBLab-VoxRex-Wav2vec2-large-CV8-da

Model description

This model is a fine-tuned version of the Swedish acoustic model KBLab/wav2vec2-large-voxrex on the Danish part of Common Voice 8.0, containing ~6 crowdsourced hours of read-aloud Danish speech.

Performance

The model achieves the following WER scores (lower is better):

Dataset	WER without LM	WER with 5-gram LM
Danish part of Common Voice 8.0	37.63	30.51
Alvenir test set	35.75	28.33