Adds a docker-compose file to set up an mlflow server to run on remote hardware. Uses additional images, MySQL for mlflow logging and minio for artefacts storage. Also included nginx which can be used for user authentication once the production version is deployed.
Please let me know what you think in its current state., the installation instructions have been tested so you should be able to get set up fine.
Some thoughts:
It might be better to have separate repos for the mlflow server and the example project. currently, they're the same repo but I feel having a separate repo for the project example would be easier to maintain and add to.
The server as it is handles everything up to training models, I've not developed the deployment/inference side of things yet. But having mlflow as our model server for that should make that straightforward.
Adds a docker-compose file to set up an mlflow server to run on remote hardware. Uses additional images, MySQL for mlflow logging and minio for artefacts storage. Also included nginx which can be used for user authentication once the production version is deployed.
Please let me know what you think in its current state., the installation instructions have been tested so you should be able to get set up fine.
Some thoughts: