Closed lauleehong closed 7 years ago
This would be a bit difficult since the base code is in lua, but a http api that exposes the model via making JSON API calls would be a good way to interface with a microphone where the code is written in python for example. A good place to start for the http api would be the waffle library here!
Hi,
I'm able to get acceptable results from short (and loud) recorded sentences, using the pre-trained DeepSpeech model trained from LibreSpeech.
Any idea how I can implement prediction.lua based on realtime speech detection using my microphone?