Use it when
- You want a framework-agnostic platform for training models, tuning hyperparameters, tracking experiments, and a built-in model registry.
- You want a platform that is opinionated and works with minimal effort.
- You want a declarative API.
- You want a platform that manages your training infrastructure with autoscaling to reduce cloud costs.
- You want to be able to deploy on-prem, major cloud providers, and Kubernetes.
Example stacks
Airflow + MLflow stack
Installation
pip install determined