README.md

metadata

language: sv-SE
datasets:
  - common_voice
  - NST Swedish ASR Database
metrics:
  - wer
tags:
  - audio
  - automatic-speech-recognition
  - speech
  - voxpopuli
license: cc-by-nc-4.0
model-index:
  - name: Wav2vec 2.0 large VoxPopuli-sv swedish
    results:
      - task:
          name: Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: NST Swedish ASR Database
        metrics:
          - name: Test WER
            type: wer
            value: 5.192353080009441
      - task:
          name: Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Common Voice
          type: common_voice
          args: sv-SE
        metrics:
          - name: Test WER
            type: wer
            value: 17.37743757973392

Wav2vec 2.0 large-voxpopuli-sv-swedish

Finetuned version of Facebooks VoxPopuli-sv large model. WER for NST + Common Voice test set (2% of total sentences) is 5.19%. WER for Common Voice test set is 17.38%.