What's wrong?

Stochastic Gradient Descent widget Regularization None cannot be selected without getting a fitting error, at least for regression (have not tried for classification).
When selecting this option, I get 'Fitting failed' error saying: 'The 'penalty' parameter of SGDRegressor must be a str among {'l2', 'elasticnet', 'l1'} or None. Got 'none' instead.'

How can we reproduce the problem?

Just try to use this type of regularization on some dataset.

What's your environment?

• Operating system: Windows 11 Enterprise
• Orange version: 3.39.0
• How you installed Orange: Downloaded and installed the software from your website.