fix ignored hmse after resource utilization computation

[irenaby]

Pull Request Description:

When GPTQ with HMSE follows resource utilization computation, it runs with MSE instead of HMSE.
Resource utilization computation runs graph_preparation_runner with default gptq flag which is false. In set_quantization_configuration_to_graph when weights_error_method is HMSE and gptq is false, weights_error_method is set to MSE in the original core_config. So when gptq is run afterwards, its config contains MSE instead of HMSE.

The fix includes:

• raise an error if HMSE is run without GPTQ (instead of overriding it)
• generate a copy of the core config and explicitly replace HMSE with MSE in resource utilization computation.

Checklist before requesting a review:

• I set the appropriate labels on the pull request.
• I have added/updated the release note draft (if necessary).
• I have updated the documentation to reflect my changes (if necessary).
• All function and files are well documented.
• All function and classes have type hints.
• There is a licenses in all file.
• The function and variable names are informative.
• I have checked for code duplications.
• I have added new unittest (if necessary).

[irenaby]

@ofirgo I'm not sure this is the best fix, opening PR so I don't forget, let's talk on Tue

[lior-dikstein]

For resource utilization, we use GPTQ disabled, and since HMSE is not supported without GPTQ we replace it with MSE.