User Guides
User Guides
Python SDK Reference

🚀 Model Operations

Model Operations features on Picsellia mainly aim at properly storing the Model and ModelVersion created with the Data Science features in the Registry and allow the user to instantiate through a Deployment from which they will be able to visualize the Prediction and ensure that the ModelVersion is actually performing as expected.

  1. Access the Model Registry and its Model Versioning System

  2. Retrieve the Experiment exported as a ModelVersion in your Private Model Registry

  3. Import your own model to your Private Model Registry

  4. Deploy a ModelVersion and monitor it with a Deployment

  5. Monitor your ModelVersion performances with a dedicated Dashboard

  6. Create a Pipeline to apply the Continuous Learning mechanism

  7. Define Alerts to be notified when something is going wrong with a Deployment

  8. Log properly the result of inferences done by the deployed ModelVersion

  9. Visualize easily the PredictedAsset and the associated Prediction generated by the deployed ModelVersion

  10. Review Prediction to send them in your Pipeline