wangbo/ComfyUI
1
0
Fork 0
mirror of https://github.com/comfyanonymous/ComfyUI.git synced 2026-02-17 08:52:34 +08:00
Code Issues Actions 16 Packages Projects Releases Wiki Activity

ops: Fix vanilla-fp8 loaded lora quality

Browse Source 
This was missing the stochastic rounding required for fp8 downcast
to be consistent with model_patcher.patch_weight_to_device.

Missed in testing as I spend too much time with quantized tensors
and overlooked the simpler ones.
This commit is contained in:
Rattus 2026-02-10 22:02:00 +10:00
parent c1b63a7e78
commit 639be0b937
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
comfy/ops.py
View File

@ -169,8 +169,8 @@ def cast_bias_weight_with_vbar(s, dtype, device, bias_dtype, non_blocking, compu
if orig.dtype == dtype and len(fns) == 0:
#The layer actually wants our freshly saved QT
x = y
else:
y = x
elif update_weight:
y = comfy.float.stochastic_rounding(x, orig.dtype, seed = comfy.utils.string_to_seed(s.seed_key))
if update_weight:
orig.copy_(y)
for f in fns: