[XPU] Fix fp8 UT patch func

[YangKai0616]

What does this PR do?

When testing tests/quantization/finegrained_fp8/test_fp8.py on XPU, it reported the following error:

FAILED tests/quantization/finegrained_fp8/test_fp8.py::FP8QuantizerTest::test_dequantization_no_accelerator - AssertionError: False is not true
FAILED tests/quantization/finegrained_fp8/test_fp8.py::FP8QuantizerTest::test_dequantize_when_no_accelerator - ValueError: Expected a xpu device, but got: cpu
FAILED tests/quantization/finegrained_fp8/test_fp8.py::FP8QuantizerTest::test_quantizer_validation_no_accelerator - AssertionError: RuntimeError not raised

The root cause is that the _patch_no_accelerator method does not fully take effect on XPU.

[YangKai0616]

Hi @MekkCyber , please help review, thanks!

[MekkCyber]

Thanks for the fix @YangKai0616

[MekkCyber]

weird that we need to add it, but I guess it's fine

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[github-actions]

[For maintainers] Suggested jobs to run (before merge)

run-slow: finegrained_fp8