Patience

[htjb]

Description

Adding a patience parameter to the MAF and clusterMAF training.

Fixes #55

Checklist:

• I have performed a self-review of my own code
• New and existing unit tests pass locally with my changes (python -m pytest)
• I have added tests that prove my fix is effective or that my feature works
• I have appropriately incremented the semantic version number in both README.rst and margarine/_version.py

[Copilot]

Pull request overview

This PR adds a configurable patience parameter to control early stopping behavior in MAF and clusterMAF training, addressing issue #55. Previously, patience was hard-coded to 2% of total epochs; now users can customize this value or use the default.

Key changes:

• Added patience parameter to MAF.train() and clusterMAF.train() methods with default of 2% of epochs
• Refactored early stopping logic to use the configurable patience value
• Updated tests to explicitly specify patience value
• Incremented version from 1.4.0 to 1.4.1

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
margarine/maf.py	Added patience parameter to train() method signature, default calculation logic, and passed to _training() method
margarine/clustered.py	Added patience parameter to train() method and forwarded to individual MAF flows
tests/test_stats.py	Updated MAF training call to include explicit patience=20
tests/test_clusters.py	Updated clusterMAF training call to include explicit patience=20
pyproject.toml	Version bump to 1.4.1
margarine/_version.py	Version bump to 1.4.1
README.rst	Version bump to 1.4.1

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.