Fix MOE tests

[ksivaman]

Description

The code drop for 2.0 results in breaking certain tests in the CI. This PR fixes the MOE/permutation tests.

Type of change

• Documentation change (change only to the documentation, either a fix or a new content)
• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Infra/Build change
• Code refactoring

Checklist:

• I have read and followed the contributing guidelines
• The functionality is complete
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

[timmoon10]

I prefer naming this quantize so that we can generalize to the other quantized tensor classes.

[ksivaman]

This is meant as a standalone quantization for a high precision tensor and I think it would be better for this to not conflict with existing quantize function in QuantizedTensor.

[ptrendx]

Wait wait wait. Why are we adding the to_float8 function again? We should just fix the tests to create and use the quantizer instead.

[yaox12]

I think the to_float8 issues has been fixed in #1468 using dequantize. Can you review that PR?

[ksivaman]

@ptrendx The to_float8 function is meant purely as a utility/convenience function for one-time/throwaway quantization use cases that just creates a defaults quantizer and returns quantized tensor. Either way, closing as this change has been included in #1468