Skip to content

vectice/validation-notebooks-repo

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

23 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Model Validation Notebooks

This repository contains a set of predefined validation notebooks to help you validate models based on best practices, with metadata and configuration files to streamline the process. You can get a prepopulated configuration files (JSON) for the model you would like to validate from the Vectice UI.

How to Use This Repository

  1. Identify which tests to run
    Choose between the available subfolders based on your specific validation needs. This repository currently includes:

    • External library tests (wrapping the MoDeVa extension)
    • Legacy tests The current folder structure separates External library tests and Legacy tests for simplicity, highlighting the difference between modern external approaches (e.g., MoDeVa extension) and legacy methods.

    While repositories can be organized differently, this setup is designed for easier navigation and understanding. In the future, the structure may evolve to group tests by type—such as Classification, Explainability, Resilience, Robustness, or Model Performance—allowing users to quickly find relevant tests.

For now, explore the folders to select the tests that best meet your validation needs.

  1. Download the Recommended Notebook
    In each subfolder, you'll find notebooks pre-configured to handle specific validation tasks. Download the notebook that best fits your requirements.

  2. Customize as Needed
    If the default notebook doesn't fully meet your needs, you can modify the provided tests or create your own. Refer to the Legacy tests folder for examples on how to add and customize your validation tests.


Folder Overview

1. External library tests

This folder wraps the MoDeVa commercial extension of PiML and provides validation notebooks for classification models. This notebook is a sample that includes tests to evaluate:

  • Explainability: Ensures that the model's predictions can be interpreted.
  • Resilience: Tests the model's ability to maintain performance under varying conditions.
  • Robustness: Measures how sensitive the model is to small changes in the input.
  • Model Performance: Evaluates traditional classification metrics like accuracy, precision, recall, F1-score, and AUC.

2. Legacy tests

The Legacy tests folder contains an example notebook that demonstrates how to add legacy validation tests to your workflow. If you wish to modify the default notebooks or create new ones, follow this guide to integrate your own tests.


How to Customize Your Notebooks

  1. Start with a Default Notebook
    Download a default notebook from Vectice APP or get one from one of the subfolders.

  2. Refer to the Custom Extensions Folder
    The notebook in custom-extensions provides examples of how to create and integrate new validation tests.

  3. Add Your Own Tests
    Follow the structure in the example to insert new tests into the existing notebook.

  4. Run Your Custom Tests
    Once modified, run your notebook with the configuration file (JSON) to validate your model with the new tests.


Feel free to explore the folders and use or modify the notebooks as needed. If you have any questions or need further assistance, reach out to the support team.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 3

  •  
  •  
  •