krita Ai - My GTX 1070 settings, Tips on getting it running faster

[@streamtabulous]

Krita Ai diffusion discussion: github.com/Acly/krita-ai-diffusion/discussions

[@streamtabulous]

oh dear bad spelling mistakes in prompt joy of my dyslexic and spelling issues, sorry to those triggered by that stuff.

[@chapicer]

valuable hints, krita ai difusion shold implement after detailer

[@streamtabulous]

you could request the features on the git hub, i haven't heard of after detailer, is that like upscale where its adding more details?

[@ikvzo]

How to fix can't open safetensor but ckpt file in krita? thx

[@BilboHobbit-b1v]

I know you have Nvidia graphic but could you by any means do some tutorial how to run the Krita AI on AMD graphics in Win11? So far I was unable to make it work on my new laptop which has AMD graphic (RX 6650M) and its such a shame :( I would be happy for any hint...

[@challengedollar2880]

yes that would be nice, following this.

[@leucome]

So search for instruction about running Comfy UI with a direct ML installation. Once you get the comfyui server working then same thing should work for Krita as it use a comfyui server.

[@DavidDoyleOutdoors]

640x640 or 704x704 may work better than 680x680 as stable diffusion likes to work in multiples of 64, you might see less artefacts

[@streamtabulous]

good to know, thank you will do that

[@challengedollar2880]

Hi there, could you please do one for GPUs above 470x , I'm rocking a 570 with 8gb but still I think its crashing in video memory or something, would be helpful to know.

[@challengedollar2880]

Hey Streamtabulous, I wanted to ask: Usually there is an "output" folder for generated images using stable diffusion, that keeps on taking memory over time, is that the case with this krita plugin? coz I would like my generated images to be only taking space the ones I use not all of them that were generated but not used.

[@streamtabulous]

not with krita, you could raise a request on aclys git hub

[@Drew_pew_pew_pew]

Strange, I have a 1070 gtx and models are running from a ssd. Copied all your settings and my generation times are much higher (02:30 for one image)

[@streamtabulous]

thats odd do you have cuda tool kit installed?

[@Drew_pew_pew_pew]

​@@streamtabulous Yes, I have. The silly thing is, I messed around a bit and managed to reduce the generation time from 14s/it to 7s/it, but I can no longer find what setting made the difference. So, an image generates in around 1'20" now. Changing the resolution from 640x640 to 1024x1024 has almost no impact on speed for me. What I did learn is that I can manually start the ComfyUI server in a terminal that Krita AI installed. This makes debugging and experimenting much easier. The next step is following your guide on installing extra samplers.

[@streamtabulous]

@@Drew_pew_pew_pew i did manually install python etc that might make a little difference. cuda_11.8.0_522.06_windows python-3.10.6-amd64 vs_BuildTools these are my installations,

[@balsalmalberto8086]

I wasn't aware of this plugin for krita.. my pc a bit better specs but it's AMD. :(

[@streamtabulous]

works with amd just set to MLC not cuda

[@balsalmalberto8086]

​@@streamtabulous I only got cuda or cpu option. I don't know if this is because I have flatpak version of krita. I have 5800xt/6800xt.

[@streamtabulous]

@@balsalmalberto8086 that's odd the should be MLC in the drop down where launch / running / install is to change from cuda. I'm not familiar with flat pack so could be a issue but not entirely sure.