Monitor, debug, and deploy

Track experiments, inspect logs, schedule production GPU workloads, and serve the models you train on AI Runtime.

Topic Description
Track runs, logs, and metrics Track experiments, metrics, and artifacts with MLflow, and view logs and GPU resource usage.
Schedule GPU workloads Deploy training code with Declarative Automation Bundles, schedule runs, and build multi-task GPU and CPU workflows.
Convert a workload to a bundle Convert an AI Runtime workload YAML into Declarative Automation Bundles to manage it as a persistent job.
Configure bundle tasks Package training code, select GPU compute and environments, and pass configuration into bundle tasks.
Serve trained models Log, register, and deploy models trained on AI Runtime to Model Serving endpoints on CPU or GPU compute.
Serve with your notebook environment Register a model from an AI Runtime notebook as an express deployment, so the endpoint uses the same environment and packages as the notebook.
Serve LLMs Serve an LLM trained on AI Runtime on a GPU endpoint behind an OpenAI-compatible API.