[Bug] 加载int-4-Dev时没任何报错直接断开Comfyui

[lestersssss]

Checklist

• 1. I have searched for related issues and FAQs (Nunchaku Frequently Asked Questions nunchaku#262) but was unable to find a solution.
• 2. The issue persists in the latest version.
• 3. Please note that without environment information and a minimal reproducible example, it will be difficult for us to reproduce and address the issue, which may delay our response.
• 4. If your report is a question rather than a bug, please submit it as a discussion at https://github.com/mit-han-lab/ComfyUI-nunchaku/discussions/new/choose. Otherwise, this issue will be closed.
• 5. I will do my best to describe the issue in English.

Describe the Bug

got promptUsing xformers attention in VAEUsing xformers attention in VAEVAE load device: cuda:0,offload device: cpu, dtype: torch.bfloat16GPU O(NVIDIA GeForce RTX4060)Memory:8187.5 MiBVRAM < 14GiB,enable CPU offload
D:\ComfyUI>pause请按任意键继续.

Environment

python3.12.9 torch:2.6.0+cu126

Reproduction Steps

[2559923437]

I encountered the same problem, but the same environment and integration package only occurred on the 40 series graphics card (RTX4070). My other graphics card devices did not have this problem. Maybe this is an error caused by the compatibility of the graphics card architecture? (RTX4070×//RTX3060&RTX3090√)

[edisonchan]

What workflow are you using?

[2559923437]

我遇到了同样的问题，但同样的环境和集成包只出现在40系显卡（RTX4070）上。我的其他显卡设备没有出现这个问题。请问这是显卡架构兼容性导致的错误吗？（RTX4070×//RTX3060&RTX3090√）

I am very sorry to ask such a question. The real reason is that I did not update the driver of the 40 series graphics card to the latest version (originally: March ×//now: May √) (but the 30 series graphics card does not seem to have such a problem)

[Nomination-NRB]

I also encounter the same problem:

VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
GPU 0 (NVIDIA GeForce RTX 3090) Memory: 24252.6875 MiB
VRAM > 14GiB，disable CPU offload
段错误 (核心已转储)

but I can't analysis the core file, when I try to use gdb, it just stop there:

got prompt
Using pytorch attention in VAE
Using pytorch attention in VAE
VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
GPU 0 (NVIDIA GeForce RTX 3090) Memory: 24252.6875 MiB
VRAM > 14GiB，disable CPU offload
段错误 (核心已转储)

(comfyui_test) ps@ps-Super-Server:/home/huishi/workspace/data/repository/ComfyUI$ gdb /home/ps/anaconda3/envs/comfyui_test/bin/python core
GNU gdb (Ubuntu 9.2-0ubuntu1~20.04.1) 9.2
Copyright (C) 2020 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Type "show copying" and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
    <http://www.gnu.org/software/gdb/documentation/>.

For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from /home/ps/anaconda3/envs/comfyui_test/bin/python...