You may want to experiment with faster-whisper: https://github.com/SYSTRAN/faster-whisper
Not sure whether or not it is fast enough on a Pi to use for real-time audio though.
Have only played with it on amd64.
Pros: pretty good with foreign languages
Cons: hallucinates on silence/background noise
Not sure whether or not it is fast enough on a Pi to use for real-time audio though.
Have only played with it on amd64.
Pros: pretty good with foreign languages
Cons: hallucinates on silence/background noise
Statistics: Posted by incognitum — Sat May 17, 2025 7:13 pm