MLOps Zoomcamp journey
Notes, homeworks, and projects for MLOps Zoomcamp.
Links
- Notes
- hw02: using MLflow to set experiment tracking with a data backend, setting tags, logging parameters; Tracking example with hyperparameter optimization using optuna here; Programmatic model registry as shown here.
- hw03: using Prefect to orchestrate flows and tasks programmatically. basically use prefect @flow and @task decorators to define an orchestration entrypoint. Now a UI or other ways e.g. CLI can be used to define a workpool to be used as compute nodes to run different jobs for example periodically run a task. Also can be used to send results via email, but many other orchestration options available as part of the prefect package. Minimal code example can be found in here, make sure to have additional setups already running such as backend server (local or remote), a pool of workers, and can populate tasks to be run.
- hw04