Skip to content

fix nvfp4 serialization #3140

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
vkuzo merged 2 commits into from
Oct 10, 2025
Merged

fix nvfp4 serialization #3140

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
fabb11d
Update
vkuzo Oct 10, 2025
fa786d1
Update
vkuzo Oct 10, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 10 additions & 0 deletions test/prototype/mx_formats/test_inference_workflow.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -191,6 +191,16 @@ def test_inference_workflow_nvfp4(
f"Got a sqnr of {sqnr} for NVFP4 recipe with bias={bias}, mm_config={mm_config}"
)

# serialization
with tempfile.NamedTemporaryFile() as f:
torch.save(m_mx.state_dict(), f)
f.seek(0)

# temporary workaround for https://github.com/pytorch/ao/issues/3077
torch.serialization.add_safe_globals([getattr])

_ = torch.load(f, weights_only=True)


class VLLMIntegrationTestCase(TorchAOIntegrationTestCase):
@pytest.mark.skipif(not torch.cuda.is_available(), reason="CUDA not available")
Expand Down
1 change: 1 addition & 0 deletions torchao/prototype/mx_formats/inference_workflow.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -211,6 +211,7 @@ def _nvfp4_inference_linear_transform(
NVFP4MMConfig,
MXGemmKernelChoice,
QuantizeTensorToMXKwargs,
QuantizeTensorToNVFP4Kwargs,
ScaleCalculationMode,
]
)
Expand Down
Loading