PyTorch Mobile - JSON model

I have a prediction model in the json format made using pytorch. To use this model, I need to catch the features of the sound (in this case we are classifying sounds) and then use the model to classify it.

I have to insert this model in an app to do this classification with the sounds that the mobile catches.

I tried to use pytorch mobile, which uses android studio, but i understood that the uploaded model must be in .pt extension and later converted to .ptl, and mine is in .json .

Is it possible to just use the json model without converting it in any other format or what steps i must follow to convert the format of this model?