[Test] Add YOLOv5 test file

[Jagggged]

• Add YOLOv5 test case for torch v2.8
• Update Dockerfile to install required packages for YOLOv5
• Add CPU fallback routine for convolution_overrideable operator

[YWHyuk]

Thanks for PR @Jagggged

The compilation error is caused by an excessive number of graph breaks.
Once the predefined limit is reached, the model start to be executed in eager mode.
This is why aten.convolution ~ was called.

I think increasing these limits will resolve the issue.

torch._dynamo.config.recompile_limit = 64
torch._dynamo.config.cache_size_limit = 128

Can you add the above code in the test file and remove aten.fallback code in PyTorchSimDevice?

[Jagggged]

As suggested, I added the two lines above, removed the fallback routine, and recommitted the changes.
Thank you for your advice!