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

mm: Remove Aimdo exemption for empty_cache (#12260)

Browse Source 
Its more important to get the torch caching allocator GC up and running
than supporting the pyt2.7 bug. Switch it on.

Defeature dynamic_vram + pyt2.7.
This commit is contained in:
rattus
2026-02-03 18:39:19 -08:00
committed by GitHub
parent fe2511468d
commit 855849c658
2 changed files with 8 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
comfy/model_management.py
View File

@@ -1724,11 +1724,9 @@ def soft_empty_cache(force=False):
elif is_mlu():
torch.mlu.empty_cache()
elif torch.cuda.is_available():
if comfy.memory_management.aimdo_allocator is None:
#Pytorch 2.7 and earlier crashes if you try and empty_cache when mempools exist
torch.cuda.synchronize()
torch.cuda.empty_cache()
torch.cuda.ipc_collect()
torch.cuda.synchronize()
torch.cuda.empty_cache()
torch.cuda.ipc_collect()
def unload_all_models():
free_memory(1e30, get_torch_device())