Example: Debug training runs

[LeoRoccoBreedt]

Description

Include a summary of the changes and the related issue.

Related to: <ClickUp/JIRA task name>

Any expected test failures?
Yes, Python versions 3.13 will fail due to PyTorch not being compatible with this version yet.

---

Add a [X] to relevant checklist items

❔ This change

• adds a new feature
• fixes breaking code
• is cosmetic (refactoring/reformatting)

---

✔️ Pre-merge checklist

• Refactored code (sourcery)
• Tested code locally
• Precommit installed and run before pushing changes
• Added code to GitHub tests (notebooks, scripts)
• Updated GitHub README
• Updated the projects overview page on Notion

---

🧪 Test Configuration

• OS: Windows
• Python version: 3.11
• Neptune version: neptune-scale 0.13.0
• Affected libraries with version: torch

Summary by Sourcery

Introduce a new “Debug training runs” example for Neptune that demonstrates tracking layer-wise gradient norms to diagnose model training issues, and integrate it into documentation and CI testing.

New Features:

• Add a tutorial notebook for debugging model training runs with Neptune
• Integrate the debug training runs example into the main README and example table

Documentation:

• Update README with links and placeholders for the new debug training runs example

Tests:

• Add the debug_training_runs notebook to the test-notebooks CI workflow

Summary by Sourcery

Add a complete “Debug training runs” example for Neptune, encompassing notebook and script versions, update documentation to reference the new tutorial, and extend CI workflows to test the new example

New Features:

• Introduce a new “Debug training runs” tutorial demonstrating layer-wise gradient norm tracking with Neptune
• Provide both a Jupyter notebook and a Python script implementation of the debug training runs example

CI:

• Include the debug training runs notebook and script in the GitHub Actions test-notebooks and test-scripts workflows

Documentation:

• Update the README to include the debug training runs example with placeholder links to documentation and projects

[sourcery-ai]

Hey @LeoRoccoBreedt - I've reviewed your changes - here's some feedback:

• The README still has TODO placeholders for [debug] and [debug-example] links—please update these to point at the final docs and Neptune project URLs before merging.
• The num_layers parameter in params isn’t actually used when building the SimpleModel—either wire it into your layer loop or remove it to avoid confusion.
• The run_examples.sh script installs requirements.txt from the current directory—consider using a relative path (e.g. -r $(dirname "$0")/requirements.txt) to ensure it runs correctly in CI.

Here's what I looked at during the review

• 🟡 General issues: 7 issues found
• 🟢 Security: all looks good
• 🟢 Testing: all looks good
• 🟡 Complexity: 1 issue found
• 🟢 Documentation: all looks good

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[sourcery-ai]

issue: Placeholder for documentation link needs to be resolved.

Replace the 'TODO' with the correct documentation link for debugging.

[SiddhantSadangi]

@LeoRoccoBreedt - I am unsubscribing from this. Please let me know once it is ready for review