2 ms·
What are the state of the art open solutions to local voice recognition? Preferably with available models that a small org can also train themselves without mi
by colechristensen 5y ago
What are the state of the art open solutions to local voice recognition? Preferably with available models that a small org can also train themselves without millions in hardware.
- albertzeyer 5y agoKaldi, K2, ESPNet.
- misc1234 5y agoNemo
- rileyphone 5y agoI will add https://github.com/coqui-ai/STT https://github.com/coqui-ai/STT, which is a continuation of DeepSpeech. Also, I've been messing around with https://github.com/ideasman42/nerd-dictation https://github.com/ideasman42/nerd-dictation, which works on a VOSK backend - accuracy is decent, especially with the bigger model.
- trowngon 5y agoVosk https://github.com/alphacep/vosk-api https://github.com/alphacep/vosk-api