I have gone through a couple of pages looking at my options.I intend to build a Pytorch model, I already have a web app all written in meteor+react. I am not sure if there is a way to combine flask and meteor. Else, is how easy is it to deploy my model for both torch and TF, then use Tensorflow.js with my react +meteor app.
Would appreciate any advice. i have most of my front end build already so that stack can’t change now.
Hope you’re going good.
Given you already have a reactjs frontend, you probably want to maintain the separation of Backend and frontend. I think in your case, it would make sense in your case to build a Flask API, and have your reactjs app simply call your external API?
I wrote an answer back here, in Sep 2019 pointing to links. Since then, PyTorch has implemented TorchServing which could be of interest - blog here, docs here.