TransformerEngine icon indicating copy to clipboard operation
TransformerEngine copied to clipboard

[PyTorch Debug] NVFP4 debug stats support

Open pggPL opened this issue 5 months ago • 1 comments

Description

This PR adds support for NVFP4 statistics: underflows and mse. I add them in seperate feature, because we may want to have a lot nvfp4-specific features added later.

Also, I renamed few variables from "fp8"-like to "quantization"-like. I cannot rename all of them - for example "is_fp8_gemm_enabled" which is an API call, so I left some of them.

Fixes # (issue)

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)
  • [x] 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:

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

pggPL avatar Oct 23 '25 14:10 pggPL

/te-ci pytorch

pggPL avatar Oct 23 '25 17:10 pggPL