mirror of
https://github.com/Comfy-Org/desktop.git
synced 2026-11-04 23:47:43 +00:00
[GH-ISSUE #1492] The outdated PyTorch version that was automatically installed for the first time on the Mac version of Desktop resulted in some functions of the k-sampler not supporting MPS. #436
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#436
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 @Nuo-Mituan on GitHub (Dec 20, 2025).
Original GitHub issue: https://github.com/Comfy-Org/desktop/issues/1492
When I was running the workflow to the k-sampler through Desktop, I found an error in aten:: nlinalg.solve-ex.result. After checking, I found out that my Mac version of Desktop automatically installed pytorch version 2.5.1 for the first time, but now the latest stable version of pytorch is 2.9.1. After trying to update it, the error in aten:: nlinalg.solve-ex.result no longer exists
Why can't the Desktop version automatically install the latest stable version of Pytorch when it first starts
┆Issue is synchronized with this Notion page by Unito
@dosubot[bot] commented on GitHub (Dec 20, 2025):
Hi @Nuo-Mituan! I'm Dosu and I’m helping the desktop team.
The Desktop app for ComfyUI on Mac installs PyTorch version 2.5.1 by default because it uses a pinned requirements file (
macos.compiled) that explicitly specifies torch==2.5.1 and related packages. This pinning is done to ensure stability and compatibility across platforms, especially given the complexity of PyTorch and MPS support on Mac. The Desktop app only updates dependencies like PyTorch when a new desktop release is published, not automatically on first launch. If you need the latest PyTorch version, you can manually update it in the.venvenvironment after installation, or use the portable/manual installation method for more control over dependencies. This approach is common for desktop apps that bundle complex dependencies, and the process may change as the app matures.Source: macos.compiled requirements file
Source: PR #1488
Source: VirtualEnvironment installer logic
To reply, just mention @dosu.
Share context across your team and agents. Try Dosu.
@Nuo-Mituan commented on GitHub (Dec 20, 2025):
I did experience a situation where Desktop couldn't open properly after updating PyTorch through (.venv), but after a while, Desktop automatically fixed it and installed the corresponding version of dependencies based on the updated PyTorch
The PyTorch team has been working on adapting MPS
Keeping the pytorch version updated on Desktop may be a good way to reduce some users giving up using Desktop due to version incompatibility
@dosubot[bot] commented on GitHub (Mar 21, 2026):
Hi, @Nuo-Mituan. 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!