wangbo/ComfyUI
1
0
Fork 0
mirror of https://github.com/comfyanonymous/ComfyUI.git synced 2026-01-10 14:20:49 +08:00
Code Issues Actions 11 Packages Projects Releases Wiki Activity

Update command line arguments in comfyui-n.bat

Browse Source 
Added '--disable-pinned-memory' to command line arguments to address recent generation issues.

removed flash-attention arguments.
This commit is contained in:
patientx 2025-12-09 16:04:49 +03:00 committed by GitHub
parent 53f0cc26e8
commit 6c552d07ef
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 2 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
comfyui-n.bat
View File

@ -1,8 +1,5 @@
@echo off
set "FLASH_ATTENTION_TRITON_AMD_ENABLE=TRUE"
set "FLASH_ATTENTION_TRITON_AMD_AUTOTUNE=TRUE"
set "MIOPEN_FIND_MODE=2"
set "MIOPEN_LOG_LEVEL=3"
@ -10,9 +7,10 @@ set "PYTHON=%~dp0venv\Scripts\python.exe"
set "GIT="
set "VENV_DIR=.\venv"
set "COMMANDLINE_ARGS=--auto-launch --use-quad-cross-attention --reserve-vram 0.9 --disable-async-offload"
set "COMMANDLINE_ARGS=--auto-launch --use-quad-cross-attention --reserve-vram 0.9 --disable-async-offload --disable-pinned-memory"
:: "--disable-async-offload" added because of this : ComfyUI has added async weight offloading by default, not compatible with older GPUs, try without it , if it works you are good.
:: "--disable-pinned-memory" added , because it seems to be the cause of many recent problems with generation using zluda.
:: in the comfyui-user.bat remove the dots on the line below and change the gfx1030 to your gpu's specific code.
:: you can find out about yours here, https://llvm.org/docs/AMDGPUUsage.html#processors