Model Registry

AI/ML Accelerator - MLflow is specifically built to track your ML experiments on the DNAnexus platform environment via the ML JupyterLab (another app in the AI/ML Accelerator package) environment. A license is required in order to use the AI/ML Accelerator package. For more information, please contact DNAnexus Sales via [e-mail].

Registering You Model

Once being logged, a model can be registered by the MLflow Model Registry. For more information on how MLflow Model Registry works, please see MLflow documentation. When a model is successfully registered, it will be available on the Models page and have version control.

You can register a model in the Model Registry through both GUI and API.

Register a Model though GUI

In the GUI of the MLflow Tracking Server, follow the below steps to register your MLflow model.

1. In the Experiments section, click on your experiment (e.g. 'TCGA Breast Cancer' as below). Then, select the run (model) that you would like to register.

{alt="Detailed page of an Experiment"}

Detailed page of an Experiment

2. You will be directed to the detailed page of the selected run (model). In this section, select 'Artifacts', and click on 'model'. Then, you will be able to register your run as a new registered model (by defining a new model name) or as a new version of an existing model (by selecting an existing registered model).

{alt="Detailed page of a selected run (model)"}

Detailed page of a selected run (model)

Register a Model though API

Another way to register your run (model) is through the MLflow model flavor or the MLflow Client Tracking API interface. Specifically, you can register a model either during an MLflow experiment run or after completing all your experiment runs. Below is an example of model registration through API.

# Define name of the registered model and assign the run ID
model_name = "tcga_breast_cancer_model" # Assign any name for your registered model
run_id = "xxxx" # ID generated by MLflow once you log a run
# Register the model to the MLflow Model Registry
registered_model = mlflow.register_model(model_uri=f"runs:/{run_id}/model", name=model_name)

run_id can be retrieved through the GUI of MLflow Tracking Server or directly from a session in your notebook via the start_run() method. See the example below:

with mlflow.start_run() as run:
## your ML scripts
run_id = run.info.run_id  # This is your run_id

Resources

Full Documentation

To create a support ticket if there are technical issues:

1. Go to the Help header (same section where Projects and Tools are) inside the platform

2. Select "Contact Support"

3. Fill in the Subject and Message to submit a support ticket.

\