Replies: 8 comments 6 replies
-
You may lower the value in the GPU Weights (MB) option, for example to 2500,as my poor NVIDIA GeForce GTX 1050 Ti 4G |
Beta Was this translation helpful? Give feedback.
-
Beta Was this translation helpful? Give feedback.
-
In webui-user.bat add in argument line |
Beta Was this translation helpful? Give feedback.
-
Hello ! i use NeverOOM and it's solving the problm for me (and it's quick) |
Beta Was this translation helpful? Give feedback.
-
i also get that message on a 3060 with 12 gb. |
Beta Was this translation helpful? Give feedback.
-
I would think that's only due to possible higher sustained temps, especially memory temps, due to higher VRAM use and possible longer generation times, and only with less than optimal cooling. I don't think or would like to hope there's no other way for damage to occur just from raw processing, unless there's some really strange edge case. |
Beta Was this translation helpful? Give feedback.
-
In the old Forge (f0.0.17v1.8.0rc-latest-276-g29be1da7), I did SDXL model generation with x2 upscaling (Hires. fix) with 8GB VRAM without any problems. Now, I decided to try the latest Forge (f2.0.1v1.10.1-previous-421-g59dd981f), and even without upscaling, I get this warning. If I set the GPU Weights to 7,000 MB, the warning stops, but the speed noticeably drops. |
Beta Was this translation helpful? Give feedback.
-
I got this same problem running on a 6gb vram gpu, but what I don't understand is that when I lower the GPU weight, instead of getting more vram space for the diffusion iteration, I get less lol. Also, could this be why when I run the Flux NF4 model, it takes 9min to load 10 steps at 512x512 while the SD1.5 takes 50secs to run 40 steps? Oh and while 5.7 of my VRAM is being used to generate the image, my GPU only 3% of my gpu power is being used in total, so I suspect my GPU is not even being used at all for FLUX for some unknown reason. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
Hello,
after the update i get this warning on a 4gb gtx 1650 super
"----------------------
[Low GPU VRAM Warning] Your current GPU free memory is 926.58 MB for this diffusion iteration.
[Low GPU VRAM Warning] This number is lower than the safe value of 1536.00 MB.
[Low GPU VRAM Warning] If you continue the diffusion process, you may cause NVIDIA GPU degradation, and the speed may be extremely slow (about 10x slower).
[Low GPU VRAM Warning] To solve the problem, you can set the 'GPU Weights' (on the top of page) to a lower value.
[Low GPU VRAM Warning] If you cannot find 'GPU Weights', you can click the 'all' option in the 'UI' area on the left-top corner of the webpage.
[Low GPU VRAM Warning] If you want to take the risk of NVIDIA GPU fallback and test the 10x slower speed, you can (but are highly not recommended to) add '--disable-gpu-warning' to CMD flags to remove this warning.
----------------------"
My 2 questions are,
where do i place the '--disable-gpu-warning' string exactly?
from where does this 'you may cause NVIDIA GPU degradation' warning come from?
A card that is made for doing things wont break by actually doing them.
A car wont break down by normal operation either. Cause its made for it.
An engine wont die by operation cause it just is made for it. ( as long its lubed ofc).
IMO this warning is clueless and has no basis what so ever. Eg, it comes from no experience what so ever.
scratching head
Beta Was this translation helpful? Give feedback.
All reactions