[PyTorch] Miscellaneous fixes for FA3 FP8 attention

[cyanguwa]

Description

This PR makes a few changes to the FA3 FP8 attention path:

• Adds descale_q, descale_k and descale_v to the FA3 call. This allows custom descaling factors for q/k/v instead of 1s. This requires FA3 PR 1210.
• Fixes the transpose in the qkv_format=sbhd case when fp8_mha=true.
• Replaces the FA3 installation warning with a debug-level logging message.
• Casts FA3 output to the same type as cuDNN attention in the unit test.

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 refractor

Changes

Please list the changes introduced in this PR:

• See description above.

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

[cyanguwa]

/te-ci pytorch

[cyanguwa]

/te-ci pytorch

[cyanguwa]

FA3: pipeline 18489052

[cyanguwa]

/te-ci pytorch

[cyanguwa]

/te-ci pytorch

[cyanguwa]

FA3 pipeline 18528978