[PyTorch] Quantize distributed BF16 weights in the fused grouped MLP

[xrennvidia]

A distributed (e.g. GTP) weight whose shards are not natively quantized is re-gathered as BF16. The fused grouped-MLP op assumed gathered weights were already quantized: forward quantized the shards before the gather, and backward read _columnwise_data / the NVFP4 amax straight off the gathered tensor, which fails for a module whose per-module recipe (e.g. MXFP8 experts under a global NVFP4 recipe) does not create native-quantized params.

Gather first, then quantize whatever is still unquantized with the op's per-expert weight quantizers, mirroring module/linear.py's handling of a re-gathered BF16 distributed weight. Native-quantized shards gather as quantized tensors and pass through unchanged. The FC2 backward re-gather moves ahead of the NVFP4 global-scale computation so that path no longer reads un-gathered shards.

Description

Please include a brief summary of the changes, relevant motivation and context.

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)
• 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

Changes

Please list the changes introduced in this PR:

• Change A
• Change B

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

[greptile-apps]

Confidence Score: 5/5

[Medium impact] The PR appears safe to merge, with the requested regression coverage now present.

Summary

The PR gathers distributed weights before quantizing BF16 results with each layer’s own quantizers.

• The fused grouped MLP quantizes distributed BF16 weights after gathering them.

Review checks were limited to source inspection; tests were not run.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart LR
  A[Distributed weights] --> B[Gather weights]
  B --> C{Already quantized?}
  C -->|Yes| D[Use gathered weights unchanged]
  C -->|No| E[Quantize with layer quantizers]
  D --> F[Forward or backward GEMM]
  E --> F

Loading

_{Reviews (3) · Last reviewed commit: "[pre-commit.ci] auto fixes from pre-comm..." · Reviewed by Greptile}