How to use Vosk models from WebSocket online server?

Asked Nov 19 '21 at 07:32

Active Nov 19 '21 at 15:02

Viewed 830 times

I have been developing an android app that uses the speech recognition service but the android device has no Google app installed. For that reason, I'm using the vosk API for speech recognition but for better accuracy in speech recognition. I need to use a higher size model. Which takes a lot of space in assets. So, how can I access the vosk model without including the assets or using them from the online server directly?

Edit:-

I have seen Kaldi's WebSocket in vosk. Can this help me to use vosk from an online server(https://github.com/just-ai/aimybox-android-sdk/tree/master/kaldi-speechkit#online-mode)?. In this, they have given information about how to use WebSocket and also given an example but I am unable to understand about making a WebSocket file.

Any help regarding this is Helpful!

edited Nov 19 '21 at 15:02

asked Nov 19 '21 at 07:32

Dev Mishra

How can a Android device not have google apps? – Khaby Lame Nov 19 '21 at 09:45
actually, it's an android watch that uses android 8.0 but it doesn't use google apps (not even google play services). – Dev Mishra Nov 19 '21 at 14:48
I have seen about WebSocket in vosk. Can this help me to use vosk from an online server – Dev Mishra Nov 19 '21 at 14:51
So what you are saying is, that for better accuracy, you want a higher size model right? – Khaby Lame Nov 21 '21 at 05:49
It's not important to use vosk api. There are other speech recognition services online and better accuracies. Check this answer on quora:https://www.quora.com/What-are-the-top-ten-speech-recognition-APIs – Khaby Lame Nov 21 '21 at 05:53
thanks it's helpfull – Dev Mishra Nov 22 '21 at 21:52

How to use Vosk models from WebSocket online server?

0 Answers0