[ENH] add feature scaling support for EncoderDecoderDataModule

[PranavBhatP]

Reference Issues/PRs

fixes #1979

What does this implement/fix? Explain your changes.

Implements feature scaling for continuous targets and features in the D2 layer using scalers and target_normalizer params.

What should a reviewer concentrate their feedback on?

PR checklist

• The PR title starts with either [ENH], [MNT], [DOC], or [BUG]. [BUG] - bugfix, [MNT] - CI, test framework, [ENH] - adding or improving code, [DOC] - writing or improving documentation or docstrings.
• Added/modified tests
• Used pre-commit hooks when committing to ensure that code is compliant with hooks. Install hooks with pre-commit install.
  To run hooks independent of commit, execute pre-commit run --all-files

[codecov]

Codecov Report

❌ Patch coverage is 50.64935% with 38 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (main@d4596de). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
pytorch_forecasting/data/data_module.py	50.64%	38 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1983   +/-   ##
=======================================
  Coverage        ?   86.67%           
=======================================
  Files           ?      160           
  Lines           ?     9540           
  Branches        ?        0           
=======================================
  Hits            ?     8269           
  Misses          ?     1271           
  Partials        ?        0           

Flag	Coverage Δ
cpu	86.67% <50.64%> (?)
pytest	86.67% <50.64%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[PranavBhatP]

Tests are failing on the notebook. Seems like its an issue caused by the training script in the notebook. Will try changing accelator = "cpu". Might resolve the issue.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[phoeenniixx]

I have doubt: Wouldn't using detach again detach the tensor from the computation graph? That would again lead to the same issue?

[PranavBhatP]

As far as my knowledge of pytorch goes, I think it's a good practice to use .detach() before converting the pytorch tensor to a numpy array. Anyways, the numpy array will not track the gradients, so this won't matter.

[phoeenniixx]

I think we could look at v1 implementation? Maybe that could help us with this issue, how the sklearn scalers are handled there?

[PranavBhatP]

Yes we will need a design doc to analyse v1 implementation. This is the best way forward.

[fkiraly]

agree - it seems like one more rabbit hole...

Is there perhaps something we can learn from pytorch-tabular?