metadata

license: apache-2.0
language:
  - lv
tags:
  - Whisper
  - CTranslate2
metrics:
  - name: wer
    type: wer
    value: 21.96
pipeline_tag: automatic-speech-recognition

Latvian Whisper tiny speech recognition model

CTranslate2 converted version for use with faster-whisper as part of Home Assistant Whisper integration

Trained on combination of:

Common Voice 17, custom selection of all validated clips, max 1000 clips per speaker
Fleurs, test+train+validation

Speech recognition quality is poor, more data is needed, donate your voice on Balsu talka