[PyTorch] Add contiguous check for te_grouped_gemm

[BeingGod]

Description

The type of input and output of te_grouped_gemm is std::vector<at::Tensor>. If the tensor is not contiguous it will causes potential accuracy problem and it is hard to debug. It's a useful safe operator to check every tensor is contiguous.

Why use is_contiguous instead of contiguous?
If tensor is non-contiguous contiguous will call aten::clone to generate a new tensor. It will cause potential performance fallback and generally user is hard to find reason (maybe them need profile).

Fixes # (issue)

Type of change

• Documentation change (change only to the documentation, either a fix or a new content)
• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Infra/Build change
• Code refractor

Changes

Please list the changes introduced in this PR:

• Add contiguous check for te_grouped_gemm

Checklist:

• I have read and followed the contributing guidelines
• The functionality is complete
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

[timmoon10]

This is a good idea. Can you sign your commit to pass the DCO check?

[BeingGod]

This is a good idea. Can you sign your commit to pass the DCO check?

Done.

[timmoon10]

/te-ci pytorch