Starter for deploying fast.ai models on your own cloud server(linux for example).
This repo can be used as a starting point to deploy fast.ai models on your own server.
Make sure you have installed docker and docker-compose before following steps blow.
- cd into the directory of the repo
- run
docker build -t fastai:lesson2 . - run
mkdir models, and put your trained model file (*.pth) under this dir. NOTE: not the directory of app/models. - run
docker-compose up -d
Then you should be able to access the application on http://yourhostip:5042