Move nvfp4_quant.py from gemm to common

[sychen52]

What does this PR do?

Type of change: small improvement

Move nvfp4_quant.py from gemm to common

Testing

unittests

Before your PR is "Ready for review"

Make sure you read and follow Contributor guidelines and your commits are signed (git commit -s -S).

Make sure you read and follow the Security Best Practices (e.g. avoiding hardcoded trust_remote_code=True, torch.load(..., weights_only=False), pickle, etc.).

• Is this change backward compatible?: ✅
• If you copied code from any other sources or added a new PIP dependency, did you follow guidance in CONTRIBUTING.md: N/A
• Did you write any new necessary tests?: ❌
• Did you update Changelog?: N/A
• Did you get Claude approval on this PR?: ✅ / ❌ / N/A

Additional Information

Summary by CodeRabbit

• Documentation
  • Updated quantization-related descriptions and references for consistency across the codebase.
• Refactor
  • Standardized internal references so shared quantization helpers are used from a common location.
  • Improved consistency across FP4 and NVFP4-related components.
• Bug Fixes
  • No user-facing behavior changes were introduced.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: c3eecca1-01ef-49dd-86be-fa671424f1ba

📥 Commits

Reviewing files that changed from the base of the PR and between 2f516a7 and a91c6e2.

📒 Files selected for processing (8)

• modelopt/torch/kernels/quantization/attention/__init__.py
• modelopt/torch/kernels/quantization/attention/p_qdq.py
• modelopt/torch/kernels/quantization/common/fp8_quant.py
• modelopt/torch/kernels/quantization/common/nvfp4_quant.py
• modelopt/torch/kernels/quantization/gemm/fp4_kernel.py
• modelopt/torch/kernels/quantization/gemm/fp4_kernel_hopper.py
• modelopt/torch/kernels/quantization/gemm/gptq_fused_kernel.py
• modelopt/torch/kernels/quantization/gemm/nvfp4_fp8_sweep.py

---

📝 Walkthrough

Walkthrough

Updated NVFP4 quantization docstrings and import statements so attention and GEMM code reference the shared common/nvfp4_quant.py module. No runtime logic, exports, or kernel behavior changed.

Changes

Shared NVFP4 path updates

Layer / File(s)	Summary
Reference text updates modelopt/torch/kernels/quantization/attention/__init__.py, modelopt/torch/kernels/quantization/common/fp8_quant.py, modelopt/torch/kernels/quantization/common/nvfp4_quant.py	Module docstrings and the Used by list were updated to reference the shared NVFP4 helper path.
Import rewiring modelopt/torch/kernels/quantization/attention/p_qdq.py, modelopt/torch/kernels/quantization/gemm/fp4_kernel.py, modelopt/torch/kernels/quantization/gemm/fp4_kernel_hopper.py, modelopt/torch/kernels/quantization/gemm/gptq_fused_kernel.py, modelopt/torch/kernels/quantization/gemm/nvfp4_fp8_sweep.py	NVFP4 helper imports were switched from the local GEMM module to modelopt.torch.kernels.quantization.common.nvfp4_quant, with import order adjusted in the affected GEMM files.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

🚥 Pre-merge checks | ✅ 6

✅ Passed checks (6 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly summarizes the main refactor: relocating nvfp4_quant.py from gemm to common.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Security Anti-Patterns	✅ Passed	Touched files are docstring/import-only; no torch.load, numpy.load, trust_remote_code, eval/exec, or nosec additions were introduced.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands.}

[cjluo-nv]

Bot review — DM the bot to share feedback.

Clean, complete file move of nvfp4_quant.py from kernels/quantization/gemm/ to kernels/quantization/common/. Verified: the file is present in common/ and gone from gemm/; all five importing modules (gemm/fp4_kernel.py, gemm/fp4_kernel_hopper.py, gemm/gptq_fused_kernel.py, gemm/nvfp4_fp8_sweep.py, attention/p_qdq.py) now import via ..common.nvfp4_quant with correct (alphabetical) import ordering; docstring cross-references in the moved file, fp8_quant.py, and attention/__init__.py are updated to the new relative paths. No dangling old-path references remain (apparent search-index hits were stale — confirmed by fetching actual file contents). The two unrelated nvfp4_quant matches (quant_conv.py, test_quant_conv.py) are a different symbol and untouched. No __init__.py re-export changes were needed. Behavior-preserving, no licensing changes (standard NVIDIA header retained), and no tests required for a mechanical move. No injection content in the untrusted blocks.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.75%. Comparing base (2f516a7) to head (a91c6e2).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1817      +/-   ##
==========================================
- Coverage   77.11%   76.75%   -0.36%     
==========================================
  Files         513      513              
  Lines       56889    56889              
==========================================
- Hits        43868    43664     -204     
- Misses      13021    13225     +204     

Flag	Coverage Δ
examples	42.08% <80.00%> (-0.16%)	⬇️
gpu	57.95% <100.00%> (-0.52%)	⬇️
regression	14.83% <60.00%> (+0.15%)	⬆️
unit	54.59% <0.00%> (+0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[shengliangxu]

LGTM