add onnxscript to optional dependencies for tests

[dennisbader]

Checklist before merging this PR:

• Mentioned all issues that this PR fixes or addresses.
• Summarized the updates of this PR under Summary.
• Added an entry under Unreleased in the Changelog.

Summary

Update:
PyTorch 2.9 made changes to torch.onnx.export(): dynamo default changed from False to True which will use a modern set of translation logic.
As long as RNN module is supported, we will continue usng `dynamo=False

Old state:

• fixes missing onnxscript for optional dependency tests
• fixes incorrect number of stored files (PyTorch 2.9+ now creates both .onnx and .onnx.data files)
• there appears to be a bug in the onnxscript library:
  • it uses op.Split(self, axis=dim, num_outputs=self.shape[dim]) which returns a single SymbolicTensor that can't be iterated over
  • instead it should do something like this: op.SplitToSequence(self, axis=dim, keepdims=False) which returns a proper sequence, then used op.SequenceAt(sequence, i) to extract individual elements

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.33%. Comparing base (97f986a) to head (884b43e).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2930      +/-   ##
==========================================
- Coverage   95.40%   95.33%   -0.08%     
==========================================
  Files         146      146              
  Lines       15707    15707              
==========================================
- Hits        14985    14974      -11     
- Misses        722      733      +11     

☔ 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.