Models

Learn more about utilizing Models within Vectice.

Models Overview

Models in Vectice represent the machine learning models created and trained during the modeling process. The model's origins, hyper-parameters, and the metrics from the training or production process are stored in the Vectice UI. Models are versioned and tagged with their deployment environment within Vectice.

Model metadata captured in Vectice

Below is a list of metadata you can find in the Vectice UI after data scientists register the models and metadata from their iterative development.

Metadata	Description
Model snapshot	Highlights the number of model versions grouped by model status.
Model Status	The model status states whether the model is in Production, Staging, or Experimentation mode.
Model creation details	Creation details show when the model was created and who it was created by.
Model type	Classifies the model by model type. For example, clustering, classification, or regression.
Algorithm	The algorithm highlights the exact modeling algorithm used for the model. (i.e., Linear Regression)
Model lineage	The model lineage can showcase the dataset used for the model, the code captured from git and the model version output.
Model metrics	The model metrics captures metrics used to evaluate model performance. (i.e., MAE and RSME values)
Versions	Models with the same name as another model you already registered in Vectice. As you register models with the same name, the versions are incremented, maintaining the model's history.
Attachments	Attachments showcase files that we're registered with the model. For example, this could be notebooks, images, or excel sheets of data analysis, model results, etc. All file types are supported for model attachments and can be downloaded from the UI. However, UI previews are only available for PNG, JPEG, CSV, Notebook, and TXT files.

Models Best Practices

• Project phases that aim to train, test, or validate models should have a step to register the models. This is typically completed in the Modeling phase of a CRISP-DM project.

• Mark your most valuable iterations and assets in the UI by selecting the star next to the corresponding iteration and assets before beginning the phase review. This will make it easier for stakeholders and subject matter experts to identify the iterations and assets in review.

Select the star next to the valuable iterations before completing a phase, even without review. This allows you to identify the most valuable iterations for that phase.

How to register models to Vectice?

View our How to register models API guide for more information on registering models during iterative development.

🎉 Great! You know more about Models in Vectice. Let's proceed and set up a new project in Vectice.