biondizzle/nvfp4-megamoe-kernel
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: use quantize_activation_nvfp4 in diag

Browse Source
This commit is contained in:
biondizzle
2026-05-23 08:39:12 +00:00
parent e2f599e4af
commit 538dbb0643
1 changed files with 4 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
tests/unit/test_nvfp4_primitives.py
View File

@@ -93,10 +93,11 @@ def test_nvfp4_primitives():
print(f" a_sf (E8M0) dtype = {a_sf_e8m0.dtype}, shape = {a_sf_e8m0.shape}")
# The runner uses E4M3 or E8M0? Check what quantize actually produces
from dsv4.ops.quantize import quantize_to_nvfp4
from dsv4.ops.quantize import quantize_activation_nvfp4
x = torch.randn(4, 512, device='cuda', dtype=torch.bfloat16)
x_fp4, x_sf = quantize_to_nvfp4(x)
print(f" quantize_to_nvfp4 output: FP4 dtype={x_fp4.dtype}, SF dtype={x_sf.dtype}")
global_scale = torch.tensor(448.0, device='cuda', dtype=torch.float32)
x_fp4, x_sf = quantize_activation_nvfp4(x, global_scale)
print(f" quantize_activation_nvfp4 output: FP4 dtype={x_fp4.dtype}, SF dtype={x_sf.dtype}")
if x_sf.dtype == torch.float8_e4m3fn:
print(f" ✅ Scale factors are E4M3 — NVFP4 correct")
elif x_sf.dtype == torch.float8_e8m0fnu: