Updating the language model

alphacep / vosk-api

Offline speech recognition API for Android, iOS, Raspberry Pi and servers with Python, Java, C# and Node

Apache License 2.0

7.38k stars 1.04k forks source link

Updating the language model #1440

Open litong11451 opened 10 months ago

litong11451 commented 10 months ago

https://alphacephei.com/vosk/adaptation After I updated the language model according to this tutorial, it can only recognize the words I added, and can no longer recognize the original words in the model

Patch4747 commented 9 months ago

I'm having this issue too. Some things I've noticed:

I have a really small "training" file, so maybe that's creating a high bias towards those words?
- See "Data collection" here.
It's not only my training words that get recognized - there are few "original" words in there too (but very little).

Let me know if you got anywhere! @litong11451