Get Started: Add a model

Just as we use experiment tracking to manage model development, it is a good idea to keep a model registry to manage the lifecycle of the models we get from our experiments. With DVC, we can start tracking models using our Git repository, which will serve as the single source of truth for our model registry and enable us to trigger automated CICD workflows. DVC Studio will then provide a model registry on top of the Git repositories to manage all our models.

Log a model artifact

All you need to start is a DVC repository and the DVCLive Python library installed:

$ pip install dvclive

Use the log_artifact method to cache the model with DVC and add it to the model registry (this snippet is from the training script in our example repository):

from dvclive import Live


with Live() as live:

...

    live.log_artifact(
        str("models/model.pkl"),
        type="model",
        name="pool-segmentation",
        desc="This is a Computer Vision (CV) model that's segmenting out swimming pools from satellite images.",
        labels=["cv", "segmentation", "satellite-images", params.train.arch],
    )

artifacts:
  pool-segmentation:
    path: models/model.pkl
    type: model
    desc:
      This is a Computer Vision (CV) model that's segmenting out swimming pools
      from satellite images.
    labels:
      - cv
      - segmentation
      - satellite-images

From model to registry

If you are building your own repository, you will need to:

1. Run the script
2. Commit it and the resulting dvc.yaml file to Git
3. Push the result to your Git remote (e.g., GitHub)

If you are following our example repository then this has already been done and we can continue to the next section to find out how to set up a model registry for all your models.