andreas/ComfyUI
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

I found a case where this is needed (#10875)

Browse Source
This commit is contained in:
comfyanonymous
2025-11-25 00:23:19 -08:00
committed by GitHub
parent acfaa5c4a1
commit 015a0599d0
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
comfy/quant_ops.py
View File

@@ -405,8 +405,8 @@ class TensorCoreFP8Layout(QuantizedLayout):
tensor_scaled = tensor * (1.0 / scale).to(tensor.dtype)
# TODO: uncomment this if it's actually needed because the clamp has a small performance penality'
# lp_amax = torch.finfo(dtype).max
# torch.clamp(tensor_scaled, min=-lp_amax, max=lp_amax, out=tensor_scaled)
lp_amax = torch.finfo(dtype).max
torch.clamp(tensor_scaled, min=-lp_amax, max=lp_amax, out=tensor_scaled)
qdata = tensor_scaled.to(dtype, memory_format=torch.contiguous_format)
layout_params = {