I am building an application for assessing students’ reading skills where students read a text passage which gets recorded and then with the use of large fine-tuned wav2vec2.0 transformer model it transcribes the audio. I have written a prototype for it and is working fine on my local computer. I read the documentation for deploying the app through streamlit sharing. It’s really fast, convenient and hassle-free way of deploying applications . But in my case I have a model folder (that contains speech-to-text models for indian different languages) that goes into GB’s in size.
So here’s my question - Is there a way through which I can deploy this application using streamlit sharing? Or Is there a more convenient form of deployment for this kind of an application (applications which involves large models)?
Thank you in advance!