Add torchvision.ops.deform_conv2d and opset19-specific tests

[f-fuchs]

Summary

Add mapping for torchvision.ops.deform_conv2d to ONNX DeformConv and add OpInfo-based test coverage.

Changes

• add torchvision.ops.deform_conv2d support in onnxscript/function_libs/torch_lib/ops/vision.py
• map PyTorch padding to ONNX pads
• infer and set group and offset_group
• support optional bias and mask

Tests

Added OpInfo coverage for:

• basic case
• bias
• mask
• nonzero padding
• grouped convolution
• multiple offset groups

Files:

• tests/function_libs/torch_lib/extra_opinfo.py
• tests/function_libs/torch_lib/ops_test_data.py

Opset handling

DeformConv requires ONNX opset 19+, while the existing torch-lib full-graph suite uses opset 18.

To keep the existing suite unchanged:

• keep the global torch-lib test baseline at opset 18
• add a dedicated deform-conv full-graph test path at opset 19
• exclude deform-conv from the generic opset-18 full-graph suite

Files:

• tests/function_libs/torch_lib/ops_test.py
• tests/function_libs/torch_lib/ops_test_common.py

Validation

python -m pytest tests/function_libs/torch_lib/ops_test.py -k deform_conv2d

Passed:

• 1 passed
• 1 skipped (expected for trace_only=True)

[f-fuchs]

@microsoft-github-policy-service agree

[justinchuby]

Thanks! For testing, I think you can instead add tests to the e2e file. This may be the simplest with regards to the opset situation

[f-fuchs]

I reverted my test changes and added two tests to tests/function_libs/torch_lib/e2e_ops_tests.py.
These then showed that my prior mapping was not correct so I fixed it. Let me know if I should add additional tests

[codecov]

Codecov Report

❌ Patch coverage is 83.33333% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 72.64%. Comparing base (8fdb1e0) to head (a45eaca).
⚠️ Report is 2 commits behind head on main.
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
onnxscript/function_libs/torch_lib/ops/vision.py	83.33%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #2923   +/-   ##
=======================================
  Coverage   72.64%   72.64%           
=======================================
  Files         259      259           
  Lines       31652    31657    +5     
  Branches     2980     2980           
=======================================
+ Hits        22994    22998    +4     
- Misses       7649     7650    +1     
  Partials     1009     1009           

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

[justinchuby]

Thanks!

[Copilot]

Pull request overview

This PR adds ONNX export support for torchvision.ops.deform_conv2d by registering a torch-lib mapping to the ONNX DeformConv operator (opset 19) and adds end-to-end export tests exercising the new path.

Changes:

• Register torchvision::deform_conv2d in torch-lib and lower it to opset19.DeformConv with stride/dilation/pad/group handling.
• Add e2e torch.onnx.export(..., opset_version=19, dynamo=True) tests for deform-conv (basic + mask/padding/group coverage).
• Update test module imports to include torchvision for the new test cases.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File	Description
tests/function_libs/torch_lib/e2e_ops_tests.py	Adds opset-19 e2e export tests for torchvision.ops.deform_conv2d.
onnxscript/function_libs/torch_lib/ops/vision.py	Registers torchvision::deform_conv2d and maps it to ONNX DeformConv (opset 19).

[justinchuby]

@f-fuchs could you follow the copilot review comments?

[f-fuchs]

@f-fuchs could you follow the copilot review comments?

done