#arc a750 taking 8 minutes to generate 30 step img in foocus 1152×896 ∣ 9:7
35 messages · Page 1 of 1 (latest)
what are you using to generate it?
foocus api
somebody mentioned about ipex is it compatible with arc a750 with foocus?
python entry_with_update.py --unet-in-bf16 --always-low-vram i use this for start sd
did you follow the instructions to get it running on Arc?
and is this windows or linux?
In the video, Bob Duffy walks us through key features and techniques for using two very different AI image generation tools, Fooocus and ComfyUI for Stable Diffusion. Both tools can run on Intel Arc GPUs and each have powerful features. However they are at the extreme ends of the user experience spectrum. So take a look to see which is best for...
ya i followed this in win 10 focuss api
i posted this on rwddit they mentioned to install py torch [i did tho]python -m pip install torch==2.1.0.post0 torchvision==0.16.0.post0 torchaudio==2.1.0.post0 intel-extension-for-pytorch==2.1.20 --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu/us/ --force-reinstall
asked there before they says they have no idea
Are you running a 5600, 5600g or 5600x?
try updating to the latest version of Pip?
yah i used this pip install numpy --upgrade
Requirement already satisfied: numpy in c:\users\systech servicing\appdata\local\programs\python\python310\lib\site-packages (1.26.4)
WARNING: You are using pip version 21.2.3; however, version 24.0 is available.
You should consider upgrading via the 'C:\Users\SYSTECH SERVICING\AppData\Local\Programs\Python\Python310\python.exe -m pip install --upgrade pip' command.
heres result
should i upgrade 2.1 to 24 does it support arc a750 focus api
Not sure tbh.
i tried updatin pip
DEPRECATION: torchsde 0.2.5 has a non-standard dependency specifier numpy>=1.19.*; python_version >= "3.7". pip 24.1 will enforce this behaviour change. A possible replacement is to upgrade to a newer version of torchsde or contact the author to suggest that they release a version with a conforming dependency specifiers. Discussion can be found at https://github.com/pypa/pip/issues/12063
@azure merlin any idea
what should i do
python entry_with_update.py --unet-in-bf16 --always-low-vram i use this to open api is it wrong way
after the command
error help me please
OSError: [WinError 126] The specified module could not be found. Error loading "C:\Users\SYSTECH SERVICING\AppData\Local\Programs\Python\Python310\lib\site-packages\torch\lib\backend_with_compiler.dll" or one of its dependencies.
how do i solve this
By the way, that problem was kind of solved after some instruction, but now it takes too long. Sometimes it takes four or five minutes to generate an image, but the first two images generate in about 45 seconds or a minute. any clue ?
I may have the same problem here, after one of the driver updates (maybe a month ago) the performance with generation degrades after each generation for some reason. Before I could generate images on lightning back to back each taking 7-8 seconds only. Now it builds up to a minute and more. In all fairness i am too lazy to revert my drivers to verify this 😄
i got this RuntimeError: Native API failed. Native API returns: -999 (Unknown PI error) -999 (Unknown PI error) error how do i solve this?
I have no idea