search

SEPIA-Framework / sepia-stt-server

SEPIA server to support open-source speech recognition via WebSocket connection.
MIT License
116 stars 21 forks source link

STT change Model #25

Open IBRI12 opened 1 month ago

IBRI12 commented 1 month ago

I followed the steps to add a vosk model. When accesing over http://localhost:20741 it is loading and understands me. But on wake up it keeps loading forever. Its the 3gb model. All installed on windows.

fquirin commented 1 month ago

That's pretty hard to debug, because I can't reproduce the problem. How much RAM does your machine have? Does the same thing happen with smaller models?

IBRI12 commented 1 month ago

I solved it deleted rnnlm and rescore its fast now.