The Verta model registry houses any model that you are iterating on.
reg_model = client.create_registered_model()
Each successive iteration is logged as a version of the registered model.
model_ver = reg_model.create_version()from verta.environment import Pythonmodel_ver.log_model(model)model_ver.log_environment(Python(["scikit-learn"]))
If you have already have a model logged to an experiment run, it can be effectively promoted to a registered model version:
model_ver = reg_model.create_version_from_run(run.id)
The Verta platform will copy the model, environment, artifacts, and applicable metadata from the experiment run to create a new model version.