I have a machine learning model for natural language processing, which is as huge as 16GB. The model used Bert and was written in python. I want to build a web api that can run this model, so that I can send the data to the model using http requests, and get the predictions as response from the website.
Currently I am using flask because it is also written in python. I tried to put the model and app.py(server file) together and push it to Heroku, but I failed because it was too large. Meanwhile, the model takes long to get the prediction out on my own computer.
Therefore, I am looking for a server/software/virtual machine that I can put my model, run it when I send my data to it, and get the response fast enough. I think possible answers might be google cloud or aws but I am not too sure.
I am really confused and if there's help I would appreciate that :)