mirror of
https://github.com/Comfy-Org/desktop.git
synced 2026-11-04 23:47:43 +00:00
[GH-ISSUE #1371] [Bug]: KSamplerAdvanced constantly got stuck at 0% (on MPS) #383
Labels
No labels
area: desktop-app
Bounty
bug
bug-cop:answered
bug-cop:response-received
component: installer
component: preload
custom node
dependencies
Design
documentation
duplicate
enhancement
expected
frontend
good first issue
installation
Low-Priority
performance
Planned
platform: cross-platform
platform: windows
Potential Bug
Potential Bug
pull-request
question
refactor
Release
stale
User Support
V1
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
Comfy-Org/desktop#383
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @kaelzhang on GitHub (Oct 11, 2025).
Original GitHub issue: https://github.com/Comfy-Org/desktop/issues/1371
App Version
System Info
The issue occurs on:
Devices
Expected Behavior
The progress of KSamplerAdvanced should increase
Actual Behavior
The progress always stuck at 0%, even after 24 hours.
Actually, when I first ran the workflow, it succeeded, the Wan2.2 model successfully generated a high-quality video from an image, but always stuck after that.
Steps to Reproduce
Just run the workflow.
Debug Logs
Browser Logs
Adding extra search path custom_nodes /Users/kael/Documents/ComfyUI/custom_nodes
Adding extra search path download_model_base /Users/kael/Documents/ComfyUI/models
Adding extra search path custom_nodes /Applications/ComfyUI.app/Contents/Resources/ComfyUI/custom_nodes
Setting output directory to: /Users/kael/Documents/ComfyUI/output
Setting input directory to: /Users/kael/Documents/ComfyUI/input
Setting user directory to: /Users/kael/Documents/ComfyUI/user
[START] Security scan
[DONE] Security scan
ComfyUI-Manager: installing dependencies done.
** ComfyUI startup time: 2025-10-11 09:18:43.723
** Platform: Darwin
** Python version: 3.12.9 (main, Feb 12 2025, 15:09:19) [Clang 19.1.6 ]
** Python executable: /Users/kael/Documents/ComfyUI/.venv/bin/python
** ComfyUI Path: /Users/kael/Documents/ComfyUI/Sources/desktop/assets/ComfyUI
** ComfyUI Base Folder Path: /Users/kael/Documents/ComfyUI/Sources/desktop/assets/ComfyUI
** User directory: /Users/kael/Documents/ComfyUI/user
** ComfyUI-Manager config path: /Users/kael/Documents/ComfyUI/user/default/ComfyUI-Manager/config.ini
** Log path: /Users/kael/Documents/ComfyUI/user/comfyui.log
[ComfyUI-Manager] Skipped fixing the 'comfyui-frontend-package' dependency because the ComfyUI is outdated.
Prestartup times for custom nodes:
0.3 seconds: /Applications/ComfyUI.app/Contents/Resources/ComfyUI/custom_nodes/ComfyUI-Manager
Checkpoint files will always be loaded safely.
Total VRAM 131072 MB, total RAM 131072 MB
pytorch version: 2.10.0.dev20250917
Mac Version (15, 4, 1)
Set vram state to: SHARED
Device: mps
Using sub quadratic optimization for attention, if you have memory or speed issues try using: --use-split-cross-attention
Python version: 3.12.9 (main, Feb 12 2025, 15:09:19) [Clang 19.1.6 ]
ComfyUI version: 0.3.64
[Prompt Server] web root: /Users/kael/Documents/ComfyUI/Sources/desktop/assets/ComfyUI/web_custom_versions/desktop_app
ComfyUI-GGUF: Allowing full torch compile
Loading: ComfyUI-Manager (V3.36)
[ComfyUI-Manager] network_mode: public
ComfyUI Revision: 1 [63722199] *DETACHED | Released on '2025-10-08'
Import times for custom nodes:
0.0 seconds: /Applications/ComfyUI.app/Contents/Resources/ComfyUI/custom_nodes/websocket_image_save.py
0.0 seconds: /Users/kael/Documents/ComfyUI/custom_nodes/ComfyUI-GGUF
0.1 seconds: /Applications/ComfyUI.app/Contents/Resources/ComfyUI/custom_nodes/ComfyUI-Manager
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/github-stats.json
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/extension-node-map.json
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/alter-list.json
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/model-list.json
Failed to initialize database. Please ensure you have installed the latest requirements. If the error persists, please report this as in future the database will be required: (sqlite3.OperationalError) unable to open database file
(Background on this error at: https://sqlalche.me/e/20/e3q8)
Starting server
To see the GUI go to: http://127.0.0.1:8000
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json
comfyui-frontend-package not found in requirements.txt
comfyui-frontend-package not found in requirements.txt
FETCH ComfyRegistry Data: 5/99
got prompt
Using split attention in VAE
Using split attention in VAE
VAE load device: mps, offload device: cpu, dtype: torch.bfloat16
Using scaled fp8: fp8 matrix mult: False, scale input: False
Requested to load WanTEModel
loaded completely 9.5367431640625e+25 6419.477203369141 True
CLIP/text encoder model load device: cpu, offload device: cpu, current: cpu, dtype: torch.float16
FETCH ComfyRegistry Data: 10/99
FETCH ComfyRegistry Data: 15/99
Requested to load WanVAE
loaded completely 9.5367431640625e+25 242.02829551696777 True
FETCH ComfyRegistry Data: 20/99
FETCH ComfyRegistry Data: 25/99
FETCH ComfyRegistry Data: 30/99
FETCH ComfyRegistry Data: 35/99
FETCH ComfyRegistry Data: 40/99
FETCH ComfyRegistry Data: 45/99
FETCH ComfyRegistry Data: 50/99
FETCH ComfyRegistry Data: 55/99
FETCH ComfyRegistry Data: 60/99
FETCH ComfyRegistry Data: 65/99
FETCH ComfyRegistry Data: 70/99
FETCH ComfyRegistry Data: 75/99
FETCH ComfyRegistry Data: 80/99
FETCH ComfyRegistry Data: 85/99
FETCH ComfyRegistry Data: 90/99
FETCH ComfyRegistry Data: 95/99
FETCH ComfyRegistry Data [DONE]
[ComfyUI-Manager] default cache updated: https://api.comfy.org/nodes
FETCH DATA from: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json [DONE]
[ComfyUI-Manager] All startup tasks have been completed.
gguf qtypes: F16 (694), Q4_K (356), Q5_K (44), F32 (1)
model weight dtype torch.float16, manual cast: None
model_type FLOW
Requested to load WAN21
0 models unloaded.
loaded completely 9.5367431640625e+25 8475.468872070312 True
0%| | 0/2 [00:00<?, ?it/s]
Settings JSON
{
"Comfy-Desktop.AutoUpdate": true,
"Comfy-Desktop.SendStatistics": false,
"Comfy.ColorPalette": "dark",
"Comfy.UseNewMenu": "Top",
"Comfy.Workflow.WorkflowTabsPosition": "Topbar",
"Comfy.Workflow.ShowMissingModelsWarning": true,
"Comfy.Server.LaunchArgs": {},
"Comfy-Desktop.UV.PythonInstallMirror": "https://github.com/astral-sh/python-build-standalone/releases/download",
"Comfy-Desktop.UV.PypiInstallMirror": "https://pypi.org/simple/",
"Comfy-Desktop.UV.TorchInstallMirror": "https://download.pytorch.org/whl/nightly/cpu",
"Comfy.InstalledVersion": "1.25.11",
"Comfy.TutorialCompleted": true,
"Comfy.Release.Version": "0.3.64",
"Comfy.Release.Status": "what's new seen",
"Comfy.Release.Timestamp": 1760142945390,
"Comfy.Minimap.Visible": false,
"Comfy.DevMode": true,
"Comfy.Locale": "en"
}
Other
No response
┆Issue is synchronized with this Notion page by Unito
@dosubot[bot] commented on GitHub (Oct 11, 2025):
Hi @kaelzhang! I'm Dosu and I’m helping the desktop team.
There are no specific fixes for KSamplerAdvanced getting stuck at 0% on MPS, but this type of issue is often related to Mac/PyTorch MPS backend limitations or incompatibilities with custom nodes. General troubleshooting steps include disabling all custom nodes to see if the issue disappears, then using binary search to isolate the problematic node if it does. If a custom node is the cause, try updating it, replacing it, or removing it if no fix is available. Users have also reported that re-initializing nodes (like re-adding a checkpoint loader) or changing the workflow structure can sometimes work around MPS-related hangs. If the issue persists without custom nodes, check your Python environment and PyTorch installation, and consider forcing CPU mode as a test. More details and troubleshooting steps are available in the official documentation and in related Mac MPS issue threads.
To reply, just mention @dosu.
Docs are dead. Just use Dosu.
@kaelzhang commented on GitHub (Oct 11, 2025):
Thanks for your reply. But it is weird that I had a successful run with the exact same workflow and image before.
@dosubot[bot] commented on GitHub (Jan 10, 2026):
Hi, @kaelzhang. I'm Dosu, and I'm helping the desktop team manage their backlog and am marking this issue as stale.
Issue Summary:
Next Steps:
Thanks for your understanding and contribution!
@kaelzhang commented on GitHub (Jan 15, 2026):
Thanks, by reducing the parameters, it works.
@dosubot[bot] commented on GitHub (Jan 15, 2026):
Thank you for your contribution and for closing this issue, kaelzhang! We appreciate your help in keeping the project tidy.