Models

Models in MarkovML represent the model binaries created during your ML workflow, which are used to predict an outcome (the target value) based on some input data (the features).

Models are created in one of two ways, depending on your use case.
When you track an ML experiment using MarkovML, we create a new Model automatically, which corresponds to the binary generated from your experiment. This Model will be associated with the experiment itself, and you will see both resources listed under the project associated with the experiment.
If instead you have a pre-existing binary which you want to evaluate using MarkovML, you can also use the Python SDK to create a new Model programmatically. Then when you record your model evaluation data with MarkovML, you can specify this model_id and the Model will be linked to the Evaluation recording. You can also reference a Model which was created as a result of running an Experiment with MarkovML, as described above.