This project is made to walk users through what model serving is, and how it fits into the machine learning workflow. This notebook will demonstrate how to connect to a model with Seldon.
To make the notebooks reproducible, we have deployed containerized notebook images on the public JupyterHub instance on the Massachusetts Open Cloud. You can get access to a Jupyter environment using your Google account! To do so, please follow the steps below:
- Visit the Operate First JupyterHub
- Click on
Log in with moc-sso
and sign in through Google. - On the spawner page, select
Image Detection
for notebook image,Large
for container size, and then clickStart server
to spawn your server. - Once your server has spawned, you should see a directory titled
pet-image-detection-<current-timestamp>
. All the notebooks should be available inside thenotebooks
directory in it for you to explore.
If you are looking for more, a version of this demo was presented at DevConf.CZ March 2021, "Beyond Inference: Bringing ML into Production." The video is available here and slides avaiable here. This talk explains the basics of model serving, why this is a relevant issue, how model serving offers relief for the data scientist/software engineer handoff, and know how to deploy a machine learning model with Seldon Core.