W&B Models Quickstart

Quickly see the mechanics for logging and linking a model to the Weights & Biases model registry:

1. run = wandb.init(): Start a run to track training
2. run.log_artifact(): Track your trained model weights as an artifact
3. run.link_artifact(): Link a specific model version it to the registry

How do you use Models in a real project?

This example keeps it simple. We're not training a real model, just focusing on the model mechanics of log_artifact() and link_artifact().

In the real world, you don't want to link every model version to the registry. Instead, use the model registry as a place to bookmark and organize your best models.

Learn more in the Models docs.