Manage Models and Projects
Last updated
Was this helpful?
Last updated
Was this helpful?
Models are a central aspect of any machine-learning application. With MarkovML, you can register your different models (organized into Projects) and track how they perform on a variety of data.
A MarkovML Model can be registered in one of two ways:
By recording a model training experiment. Tracking an experiment automatically creates and registers a Model with MarkovML. The Track Experiments guide covers experiment tracking.
By explicitly creating a model using the MarkovML SDK. This is useful when you want to record a model evaluation without tracking an experiment first. See the Record Model Evaluations guide for more details.
All Models in MarkovML and their related resources (Experiments and Evaluations) belong to a Project. Projects help you keep the various resources for your ML applications organized.