Serving ML Models as Reusable Containers

We want to create a class which contains methods to download a model from S3 and load it directly into memory. Something as simple as this should suffice for demonstration purposes. Once you've built the image you can run your container using the following command: Change the model url to your S3 URL and the docker image to point to your docker registry. I've put mine up there just as an example. Now here comes the best part.