Artifact versioning and configuration

Understand and adjust how ZenML versions your data.

Each artifact that ZenML saves in your artifact store is assigned a name and is versioned automatically. Using the ArtifactConfig class, you can annotate your ZenML steps to modify the name, version, and other properties of the artifacts created by the step:

from zenml import step, ArtifactConfig

@step
def my_step() -> Annotated[int, ArtifactConfig(name=..., version=..., ...)]:
    ...


@step
def my_multi_output_step() -> Tuple[
    Annotated[int, ArtifactConfig(...)],
    Annotated[int, ArtifactConfig(...)],
]:
    ...

Assign custom artifact names

If you do not give your outputs custom names, the artifacts created by your ZenML steps will be named {pipeline_name}::{step_name}::output or {pipeline_name}::{step_name}::output_{i} by default:

from zenml import pipeline, step

@step
def my_step() -> int:
    ...

@step
def my_multi_output_step() -> Tuple[int, int]
    ...

@pipeline
def my_pipeline():

    # Default artifact name will be `my_pipeline::my_step::output`
    my_step()

    # Default artifact names will be
    #   - `my_pipeline::my_multi_output_step::output_1`
    #   - `my_pipeline::my_multi_output_step::output_2`
    my_multi_output_step()

To give your artifacts custom names, use the name parameter of the ArtifactConfig class:

from zenml import step, ArtifactConfig

@step
def my_step() -> Annotated[int, ArtifactConfig(name="my_custom_name")]:
    ...

Now you will be able to find the created artifact as my_custom_name in your ZenML dashboard.

To prevent visual clutter, only artifacts with custom names are displayed in the ZenML dashboard by default. Thus, make sure to assign names to your most important artifacts that you would like to explore visually.

Assign custom artifact versions

ZenML automatically versions all created artifacts using auto-incremented numbering. I.e., if you have defined a step creating an artifact named "my_custom_artifact" as shown above, the first execution of the step will create an artifact with this name and version "1", the second execution will create version "2" and so on.

If you would like to override the version assigned to an artifact, you can use the version property of the ArtifactConfig:

from zenml import step, ArtifactConfig

@step
def my_step() -> (
    Annotated[
        int, 
        ArtifactConfig(
            name="my_custom_name", 
            version="my_custom_version"
        ),
    ]
):
    ...

The next execution of this step will then create an artifact with the name "my_custom_name" and version "my_custom_version". This is primarily useful if you are making a particularly important pipeline run (such as a release) whose artifacts you want to distinguish at a glance later.

You cannot create two artifacts with the same name and version, so rerunning a step that specifies a manual artifact version will always fail.

Assign tags to your artifacts

If you want to tag the artifacts of a step or pipeline that is executed repeatedly, you can use the tags property of ArtifactConfig to assign an arbitrary number of tags to the created artifacts:

from zenml import step, ArtifactConfig

@step
def my_step() -> (
    Annotated[int, ArtifactConfig(tags=["my_tag_1", "my_tag_2"])]
):
    ...

This will assign tags "my_tag_1" and "my_tag_2" to all artifacts created by this step, which can later be used to filter for artifacts in the dashboard.

Link artifacts to models

ArtifactConfig contains several other properties that you can set to modify how artifacts are linked to models:

• model_name: The name of the model to link the artifact to.

• model_version: The version of the model to link the artifact to.

• is_model_artifact: Whether the artifact is a model artifact.

• is_endpoint_artifact: Whether the artifact is an endpoint artifact.