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

Dynamic VRAM fixes - Ace 1.5 performance + a VRAM leak (#12368)

Browse Source 
* revert threaded model loader change

This change was only needed to get around the pytorch 2.7 mempool bugs,
and should have been reverted along with #12260. This fixes a different
memory leak where pytorch gets confused about cache emptying.

* load non comfy weights

* MPDynamic: Pre-generate the tensors for vbars

Apparently this is an expensive operation that slows down things.

* bump to aimdo 1.8

New features:
watermark limit feature
logging enhancements
-O2 build on linux
This commit is contained in:
rattus
2026-02-09 13:16:08 -08:00
committed by GitHub
parent a0302cc6a8
commit 62315fbb15
5 changed files with 20 additions and 35 deletions
Download Patch File Download Diff File Expand all files Collapse all files
comfy/ops.py
+1 -1
View File
@@ -87,7 +87,7 @@ def cast_bias_weight_with_vbar(s, dtype, device, bias_dtype, non_blocking, compu
signature = comfy_aimdo.model_vbar.vbar_fault(s._v)
if signature is not None:
xfer_dest = comfy_aimdo.torch.aimdo_to_tensor(s._v, device)
xfer_dest = s._v_tensor
resident = comfy_aimdo.model_vbar.vbar_signature_compare(signature, s._v_signature)
if not resident: