Add ip_workloads to checkin tests

[ssapreTT]

ip_workloads are not being exercised in checkin tests.

@ramamalladiTT
@lyen1

[ramamalladiTT]

But not all ip_workloads execute. For example, the mamba2 ONNX test fails because the onnx file is missing. The user is expected to generate the onnx file from workloads/onnx/examples.

[ssapreTT]

About ONNX files, I think we should have a type ONNXPY, similar to ONNX, which will do this - run the onnx creation python file, and then consume the generated file. That way we can consume such "to be generated" files too.

About ip_workloads, is there a reason to maintain it as a separate file? Can we not maintain all the workloads in all_workloads.yaml?

[ramamalladiTT]

The onnx creation isn't quick... so, including that in check-in tests isn't a good idea. We can possibly keep that as 2nd level tests if needed.

Yes, we can do away with ip_workloads and merge into all_workloads.