June 22, 2023

This release includes a number of major enhancements to the ValidMind Library that will make it easier for users to edit templates and add custom tests that can be reused across templates.

In addition, this release also includes a new notebook to illustrate support for NLP models and PyTorch, as well as the ability to edit templates directly in the platform.

Release highlights

ValidMind Library (v1.15.4)

Implement your own custom tests

With support for custom tests, you can now go beyond the default set of documentation and testing components provided by ValidMind, and use our library to document any type of model or use case.

Integrate external test providers

With test providers, you can now integrate external test libraries to expand the set of tests available through the ValidMind Library, or create your own test libraries.

  • This enables registering custom tests under a namespace to make them available globally.
  • We provide a couple of demo scenarios with such test providers and show how you can register them with ValidMind so that they can run a template utilizing your tests.

Simplified documentation and testing using templates

Templates now function as dynamic test suites, allowing you to identify all the documentation components and test results mapped to a specific template.

This makes it easier to fill a pre-configured template (including boilerplates and spaces designated for documentation and test results) with a single command, instead of running multiple test suites.

New notebook to illustrate natural language processing (NLP) data analysis and text classification use case

Learn how to use the ValidMind Library to document and test an NLP data analysis and text classification model. The emphasis of this notebook is on documenting and testing the in-depth analysis and preprocessing of the text data, in this case, COVID-19 related tweets.

ValidMind Platform

Editable templates

You can now edit the documentation templates that get used for model documentation or for validation reports.

  • This feature is relevant for administrators who need to configure templates for specific use cases, or where the existing templates supplied by ValidMind need to be customized.
  • Editing allows to configure sections, sub-sections, and content blocks inside a template.

Enhancements

We revised our QuickStart guide to be more modular and to highlight that our suggested starting point with the ValidMind Library is now JupyterHub.

How to upgrade

ValidMind Platform

To access the latest version of the ValidMind Platform,1 hard refresh your browser tab:

  • Windows: Ctrl + Shift + R OR Ctrl + F5
  • MacOS: ⌘ Cmd + Shift + R OR hold down ⌘ Cmd and click the Reload button

ValidMind Library

To upgrade the ValidMind Library:2

  1. In your Jupyter Notebook:

    • Using JupyterHub:3 Hard refresh your browser tab.
    • In your own developer environment:4 Restart your notebook.
  2. Then within a code cell or your terminal, run:

    %pip install --upgrade validmind

You may need to restart your kernel after running the upgrade package for changes to be applied.