Yolov4 head sub_module fails in main

[punithsekar]

Describe the bug
yolov4 head sub_module fails in main after this commit in main.

RuntimeError: TT_FATAL @ ../ttnn/cpp/ttnn/operations/matmul/device/matmul_op.cpp:977: bias_shape.without_padding()[-1] == b_shape[-1]
E       info:
E       Unsupported bias shape: last dimension not equal to second input's last dimension.
E       backtrace:
E        --- ttnn::operations::matmul::Matmul::validate(std::__1::vector<tt::tt_metal::Tensor, std::__1::allocator<tt::tt_metal::Tensor>> const&, std::__1::vector<std::__1::optional<tt::tt_metal::Tensor const>, std::__1::allocator<std::__1::optional<tt::tt_metal::Tensor const>>> const&) const 

To Reproduce
Steps to reproduce the behavior:

1. Checkout keerthanar/yolov4_bias_issue_unit_test
2. Run pytest tt-metal/tests/ttnn/unit_tests/operations/test_new_conv2d.py::test_yolov4_bias_issue_case

Please complete the following environment information:

• OS: Ubuntu 20.04
• Device- n150

[punithsekar]

Temporarily, I have skipped those test in main.

PR #11959

[tt-rkim]

Note @eyonland for planning

[mywoodstock]

cc: @bbradelTT

[bbradelTT]

Starting to work on this.

1. I think test_yolov4_bias_issue_case is duplicated in the file on the branch
2. I think the repro step should be

pytest tests/ttnn/unit_tests/operations/test_new_conv2d.py::test_yolov4_bias_issue_case

3. I can check for both padded and non-padded. That allows the test to pass. I need to see if there are any negative consequences. If not, I'll get a PR out.

[keerthana-r-mcw]

Hi @bbradelTT , The issue got resolved with this matmul_op.cpp fix given by Sankar.
NOTE: We tested the fix locally, but it has not yet been merged into the main.

[punithsekar]

The issue has been solved.
#12503 PR for removing the skip statements.

[punithsekar]

Merged #12503 to main.
Closing this issue.