ComfyUI Version 0.3.59 with Python 3.12.10 + Torch 2.7.0 + cuda 12.8
this is the official version, the only difference are the installed libs and the installed custom nodes:
Installed Libs
torch=2.7.0+cu128
insightface-0.7.3-cp312
facexlib
nvidia apex
sage attention 2.2.0
flash attention 2 (flash_attn-2.7.4)
triton-windows 3.3.1
xformer 0.0.30
numpy 1.26.4 (important for insightface)
a working pulid flux version
reactor with bypassed nsfw check
Installed nodes
blibla-comfyui-extensions
comfy-mtb
comfyui_birefnet_ll
comfyUI_bitsandbytes_NF4
ComfyUI_bnb_nf4_fp4_Loaders
ComfyUI_Comfyroll_CustomNodes
comfyui_controlnet_aux
comfyui_essentials
comfyui_essentials_mb
comfyui_fizznodes
comfyui_ipadapter_plus
comfyui_layerstyle
ComfyUI_LayerStyle_Advance
comfyui_memory_cleanup
comfyui_patches_ll
comfyui_pulid_flux_ll
comfyui_tinyterranodes
comfyui_ttp_toolset
comfyui_vlm_nodes
comfyui-advanced-controlnet
comfyui-animatediff-evolved
comfyui-art-venture
ComfyUI-Crystools
comfyui-custom-scripts
comfyui-depthanythingv2
comfyui-detail-daemon
comfyui-easy-use
comfyui-florence2
comfyui-frame-interpolation
ComfyUI-GGUF
ComfyUI-GIMM-VFI
comfyui-ic-light
comfyui-image-selector
comfyui-impact-pack
comfyui-impact-subpack
comfyui-inspire-pack
comfyui-jdcn
comfyui-kjnodes
ComfyUI-Manager
ComfyUI-Miaoshouai-Tagger
comfyui-multigpu
comfyui-mxtoolkit
comfyui-propost
comfyui-reactor
ComfyUI-RvTools_v2
ComfyUI-segment-anything-2
comfyui-superbeasts
comfyui-supir
comfyui-videohelpersuite
ComfyUI-WanVideoWrapper
efficiency-nodes-comfyui
RES4LYF
rgthree-comfy
sd-dynamic-thresholding
sdxl_prompt_styler
teacache
virtuoso-nodes
wanblockswap
was-ns
ComfyUI Version 0.3.51 with Python 3.12.10 + Torch 2.7.0 + cuda 12.8
IMPORTANT: Change back the Security Level in this file from weak to normal:
ComfyUI_windows_portable\ComfyUI\user\default\ComfyUI-Manager\config.ini
it is set to weak to install nodes from the dev channel like bitsandbytes (nf4)
security_level = weak -> security_level = normal
this is the official version, the only difference are the installed libs and the pre-installed custom nodes:
Installed Libs
insightface
facexlib
flash_attn 2
sage attn 2.2.0
xformer 0.0.30
triton-windows 3.3.1.post19
comfyui-frontend-package 1.26.0
Installed Nodes
blibla-comfyui-extensions
comfy-mtb
comfyui_birefnet_ll
ComfyUI_bitsandbytes_NF4
ComfyUI_Comfyroll_CustomNodes
comfyui_controlnet_aux
comfyui_essentials
comfyui_essentials_mb
comfyui_fizznodes
comfyui_ipadapter_plus
ComfyUI_JPS-Nodes
comfyui_layerstyle
ComfyUI_LayerStyle_Advance
comfyui_patches_ll
comfyui_pulid_flux_ll
comfyui_ttp_toolset
comfyui-advanced-controlnet
comfyui-animatediff-evolved
ComfyUI-Crystools
comfyui-custom-scripts
comfyui-detail-daemon
comfyui-easy-use
comfyui-florence2
comfyui-frame-interpolation
ComfyUI-GGUF
comfyui-image-selector
comfyui-impact-pack
comfyui-impact-subpack
comfyui-inspire-pack
comfyui-jdcn
comfyui-kjnodes
ComfyUI-Manager
ComfyUI-Miaoshouai-Tagger
comfyui-multigpu
comfyui-mxtoolkit
comfyui-post-processing-nodes
comfyui-reactor
ComfyUI-RvTools_v2
ComfyUI-segment-anything-2
comfyui-superbeasts
comfyui-supir
comfyui-videohelpersuite
ComfyUI-WanVideoWrapper
comfyui-wd14-tagger
efficiency-nodes-comfyui
pulid_comfyui
rgthree-comfy
sd-dynamic-thresholding
sdxl_prompt_styler
teacache
virtuoso-nodes
was-ns
ComfyUI Version 0.3.49 with Python 3.12.10 + Torch 2.7.0 + cuda 12.8
After the last update to 0.3.50/0.3.51, nothing worked anymore, and I mean nothing. The nodes returning type mismatches and the workflows simply stopped running. So I created a sabere version from version 0.49, installed everything that was important (to me) and created an archive of it so I wouldn't have to do it ever again.
this is the official version, the only difference are the installed libs and the pre-installed custom nodes:
Installed Libs
torch=2.7.0+cu128
insightface-0.7.3-cp312
facexlib
nvidia apex
sage attention 2.2.0
flash attention 2 (flash_attn-2.7.4)
triton-windows 3.3.1
xformer 0.0.30
numpy 1.26.4 (important for insightface)
a working pulid flux version (to fix the issues with eva clip it should be enough to install: python_embeded\python.exe -m pip install pytorch-extension)
reactor with bypassed nsfw check
Installed nodes
blibla-comfyui-extensions
comfy-mtb
comfyui_birefnet_ll
comfyUI_bitsandbytes_NF4
ComfyUI_bnb_nf4_fp4_Loaders
ComfyUI_Comfyroll_CustomNodes
comfyui_controlnet_aux
comfyui_essentials
comfyui_essentials_mb
comfyui_fizznodes
comfyui_ipadapter_plus
comfyui_layerstyle
ComfyUI_LayerStyle_Advance
comfyui_memory_cleanup
comfyui_patches_ll
comfyui_pulid_flux_ll
comfyui_tinyterranodes
comfyui_ttp_toolset
comfyui_vlm_nodes
comfyui-advanced-controlnet
comfyui-animatediff-evolved
comfyui-art-venture
ComfyUI-Crystools
comfyui-custom-scripts
comfyui-depthanythingv2
comfyui-detail-daemon
comfyui-easy-use
comfyui-florence2
comfyui-frame-interpolation
ComfyUI-GGUF
ComfyUI-GIMM-VFI
comfyui-ic-light
comfyui-image-selector
comfyui-impact-pack
comfyui-impact-subpack
comfyui-inspire-pack
comfyui-jdcn
comfyui-kjnodes
ComfyUI-Manager
ComfyUI-Miaoshouai-Tagger
comfyui-multigpu
comfyui-mxtoolkit
comfyui-propost
comfyui-reactor
ComfyUI-RvTools_v2
ComfyUI-segment-anything-2
comfyui-superbeasts
comfyui-supir
comfyui-videohelpersuite
ComfyUI-WanVideoWrapper
efficiency-nodes-comfyui
RES4LYF
rgthree-comfy
sd-dynamic-thresholding
sdxl_prompt_styler
teacache
virtuoso-nodes
wanblockswap
was-ns
Description
Re-Upload and Updated version 0.3.49_torch2.7.0_cu128
fixed an issue with the installed xformer version
fixed a typo in the run_nvidia_gpu.bat disabling xformer
also updated the comfyui settings (yea sorry i don't use the "new menu" xd)