Ways to improve performance on AMD graphics? #3940
Unanswered
someguy217896
asked this question in
Q&A
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
I'm using AMD Radeon RX 7900 GRE with has 16 GB Gpu.
I've successfully installed the program using the Windows (AMD GPUs) guide. Also I moved the files to an external SSD.
I've noticed that when loading it only uses 1024 VRAM, is there a way to make it use more?
I found this:
#2302 (comment)
Would this help? In which file and where do I paste:
EXTRA_PARAMS = [
"--always-high-vram",
"--disable-offload-from-vram",
"--disable-in-browser",
]
sys.argv = EXTRA_PARAMS
or is there a reason VRAM is capped at 1024, due to GPU limitations?
It seems to want to reload the model for each image generation, I'm guessing this is a necessary process?
Requested to load SDXL
Loading 1 new model
loading in lowvram mode 64.0
[Fooocus Model Management] Moving model(s) has taken 4.65 seconds
100%|██████████████████████████████████████████████████████████████████████████████████| 30/30 [01:36<00:00, 3.20s/it]
Requested to load AutoencoderKL
Loading 1 new model
loading in lowvram mode 64.0
[Fooocus Model Management] Moving model(s) has taken 0.33 seconds
[Fooocus] Saving image 1/2 to system ...
Image generated with private log at: D:\Fooocus_win64_2-5-0\Fooocus_win64_2-5-0\Fooocus\outputs\2025-04-02\log.html
Generating and saving time: 103.25 seconds
[Fooocus] Preparing task 2/2 ...
[Sampler] refiner_swap_method = joint
[Sampler] sigma_min = 0.0291671771556139, sigma_max = 14.614643096923828
Requested to load SDXL
Loading 1 new model
loading in lowvram mode 64.0
[Fooocus Model Management] Moving model(s) has taken 63.39 seconds
It takes a whole 60 seconds on the second attempt.
Beta Was this translation helpful? Give feedback.
All reactions