Bhashini-2024 / Bhashini-C4GT

0 stars 1 forks source link

Added ASR_app, Scripts and assets for Speech Recognition quantised models. #3

Closed aryan1165 closed 4 months ago

aryan1165 commented 4 months ago

In accordance to https://github.com/Bhashini-2024/Bhashini-C4GT/issues/1 .

In this pull request, I have added the code for Demo ASR android application. It doesn't have the model yet. Also I have added all the scripts to export from .nemo checkpoint, perform static and dynamic quantisation on these exported models and inferencing of all the models. Also the Assets directory has the skeleton of where these models can be stored.

As the model files were really heavy they are not added right now.