A tour of erm, a local CLI that removes disfluencies from English speech recordings using faster-whisper, a few detectors that look at the audio directly, and ffmpeg.
  • ultimate_worrier@lemmy.dbzer0.com
    5·
    2 days ago

    Voice to speech seems pretty non deterministic. Looks inside: openAI

    It would be hilarious if you built a cli that adds question marks when someone uptalks.