July 22, 2024

On top of our increased support for tests and arrays in the ValidMind Library and our first ValidMind Academy training modules, we've introduced customizable model workflows, new role functionality, and new report types.

Release highlights

ValidMind Library (v2.4.4)

Support for comparison tests

You can now run comparison tests with the ValidMind Library.

Run comparison tests

• This feature is useful when you want to run the same test against multiple combinations of models or datasets.
• A comparison test creates a single documentation block that compares the individual results.

Comparison test being run in a notebook

The updated run_test() function allows you to pass an input_grid which runs a test for all combinations of inputs.

Input grid example

input_grid = {
    "model": ["XGBoost"],
    "dataset": ["train_dataset", "test_dataset"],
}

A test runs once for each of the following input groups:

{
"model": "XGBoost",
"dataset": "train_dataset"
}

{
"model": "XGBoost",
"dataset": "test_dataset"
}

Example function calls

from validmind.tests import run_test

input_grid = {
    "model": ["XGBoost"],
    "dataset": ["train_dataset", "test_dataset"],
}

result = run_test(
    "validmind.model_validation.sklearn.ClassifierPerformance",
    input_grid,
)

result = run_test(
    "validmind.model_validation.sklearn.ConfusionMatrix",
    input_grid,
)

result = run_test(
    "validmind.model_validation.sklearn.ROCCurve",
    input_grid,
)

ValidMind Platform (v1.22.0)

Customizable model workflows

You can now manage lifecycle processes within your ValidMind Platform setup using customizable model workflows.

• You use these workflows to match your organizational needs for overseeing model development, validation, or implementation activities.
• A default workflow is provided for you as suggestion.

Setting up workflows

Workflow step configuration

Enhancements

ValidMind Library (v2.4.4)

Support for 2D arrays in add_extra_column

• With enhancements to the VMDataset class, the add_extra_column() method now supports adding 2D arrays as single columns in the dataset's DataFrame.
• These changes ensure that 2D arrays are stored in a single column without altering the DataFrame structure.

add_extra_column()

For example:

import numpy as np
import pandas as pd

# sample DataFrame
df = pd.DataFrame({
    'A': [1, 2, 3],
    'B': [4, 5, 6]
})

# sample 2D array
array_2d = np.array([
    [1, 2],
    [3, 4],
    [5, 6]
])

vm_dataset = vm.init_dataset(
    dataset=df,
    input_id=dummy_ds
)

vm_dataset.add_extra_columns(
    "dummy_column",
    array_2d
)

ValidMind Platform (v1.22.0)

Add new role functionality

Users with the Create Role permission can now add a new role under Settings.

Manage roles

Manage permissions

New report types

We added a number of new report types:

• Number of models by business unit and tier
• Number of models by status
• Avgerage number of days models spend in a status
• Models by number of findings

Working with analytics

Bug fixes

ValidMind Library (v2.4.4)

We fixed a number of missing test descriptions that were caused by a scripting issue.

Test descriptions

Documentation

ValidMind Academy

We're introducing the first training modules that are part of our ValidMind Academy training program for:

• Model developers
• Model validators
• Administrators

How to upgrade

ValidMind Platform

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

¹ Log in to ​ValidMind

• 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

² ValidMind Library

1. In your Jupyter Notebook:

   • Using JupyterHub: Hard refresh your browser tab.
   • In your own developer environment: 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.