Broadcast the two input shapes for transposed matmul

[FelixXidddd]

Description

There is a transposed matmul layer(cases from meta_benchmark), has two inputs src0 and src1(which is a const).
Error message:

src[0] (Unnamed Layer* 1286) [Shuffle]_output,nbExtraDims 1, [2048,752,200], op TRANSPOSE; src[1] (Unnamed Layer* 1287) [Constant]_output, [752,60], op NONE; 
[10/19/2022-01:00:22] [TRT] [E] 4: [optimizer/shapeof/graphShapeAnalyzer.cpp::analyzeShapes::1877] Error Code 4: Miscellaneous (IMatrixMultiplyLayer [MATRIX_MULTIPLY]-[acc_ops.trt_transposed_matmul]-[trt_transposed_matmul]: broadcast dimensions must be conformable)
getDimensions failed xx: (Unnamed Layer* 1288) [Matrix Multiply]_output

As you can see from the error message, the src0 is [2048,752,200] and src1 is [752,60] which is not broadcast conformable. Actually torch should unsqueeze src1 from [752,60] to [1,752,60]

Fixes # (issue)
So we need to broadcast the inputs of transposed matmul layer.

Type of change

Please delete options that are not relevant and/or add your own.

• Bug fix (non-breaking change which fixes an issue)

Checklist:

• My code follows the style guidelines of this project (You can use the linters)
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas and hacks
• I have made corresponding changes to the documentation
• I have added tests to verify my fix or my feature
• New and existing unit tests pass locally with my changes
• I have added the relevant labels to my PR in so that relevant reviewers are notified

[nvpohanh]

@ncomly-nvidia Could you let us know what we need to do to have this reviewed and merged? Thanks

[ncomly-nvidia]

@narendasan @frank-wei can you please review or assign?

[yinghai]

Thank you for the contribution. Could you add a unittest?

[FelixXidddd]

@yinghai Could you help to point out where should I add a unit test which can save me a lot of time.

[yinghai]

For example: https://github.com/pytorch/TensorRT/blob/master/py/torch_tensorrt/fx/test/converters/acc_op/test_matmul.py

[facebook-github-bot]

Hi @FelixXidddd!

Thank you for your pull request and welcome to our community.

Action Required

In order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you.

Process

In order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA.

Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with CLA signed. The tagging process may take up to 1 hour after signing. Please give it that time before contacting us about it.

If you have received this in error or have any questions, please contact us at cla@meta.com. Thanks!

[FelixXidddd]

The corresponding test is added. Please help to review.

[FelixXidddd]

@narendasan Should I sign up at https://code.facebook.com/cla? Should not we (Nvidia employees) already sign?

[frank-wei]

LGTM! Pls sign up for it since every contributor needs to do so.

[ncomly-nvidia]

@narendasan Should I sign up at https://code.facebook.com/cla? Should not we (Nvidia employees) already sign?

@FelixXidddd you need to sign the corporate one. @narendasan can you please send the info over?

[nvpohanh]

Let's use #1457 instead and close this. Thanks