Implement linalg lowering of diag_embed torch op

[schnkmwt]

This PR adds lowering of diag_embed to linalg dilect.
Tracked in nod-ai/AMD-SHARK-ModelDev#288

[ramiro050]

Thanks! Looks mostly good, I just have a few comments

[newling]

I'm still trying to understand how the indexing works. Test coverage looks good, and the comments are useful. Nice one!

[ramiro050]

Not sure if you're manually making the lint fixes, but you can use git clang-format to automatically format the changes.

git clang-format main

formats your entire branch. If you don't have git clang-format, you should be able to do inside torch-mlir/

./externals/llvm-project/clang/tools/clang-format/git-clang-format main

[schnkmwt]

Not sure if you're manually making the lint fixes, but you can use git clang-format to automatically format the changes.

git clang-format main

formats your entire branch. If you don't have git clang-format, you should be able to do inside torch-mlir/

./externals/llvm-project/clang/tools/clang-format/git-clang-format main

For some reason git clang-format main did not work for me. I guess it only works on most recent commit. But the git-clang-format under ./externals worked like a charm! Thanks for the suggestion.

[ramiro050]

LGTM, I think you need to XFAIL your tests on the ONNX backend

[schnkmwt]

LGTM, I think you need to XFAIL your tests on the ONNX backend

Yes just pushed that. Got help from discord!