3 ms·
I have a mediocre GPU but a fast CPU (with a lot of RAM). Would I see improvements there? I guess I should give it a try.
by gpderetta 4y ago
I have a mediocre GPU but a fast CPU (with a lot of RAM). Would I see improvements there?
I guess I should give it a try.
- senthilnayagam 4y agotried it yesterday, on intel i9 macbook pro it takes about 300 seconds per image.
- gpderetta 4y agoYou mean the keras version? How does it compare to the original one? Currently on my 10850k I get 2.4s/iteration, which is borderline usable. I haven't managed (nor tried very hard) to get the cuda version working on my 1070; I expect to be a little better, but I don't want to fight with ram issues.
- ttflee 4y agoHow many steps did you perform? I tried some and found no major differences after 16 steps or so with given random seed.
- ttflee 4y agoOn intel MacBookPro 2020, CPU-only, the original one[1] using pytorch utilized one core only. A tensorflow implementation[2] with oneDNN support which utilized most of the cores ran at ~11sec/iteration. Another OpenVINO based implementation[3] ran at ~6.0sec/iteration. [1] https://github.com/CompVis/stable-diffusion/ https://github.com/CompVis/stable-diffusion/ [2] https://github.com/divamgupta/stable-diffusion-tensorflow/ https://github.com/divamgupta/stable-diffusion-tensorflow/ [3] https://github.com/bes-dev/stable_diffusion.openvino/ https://github.com/bes-dev/stable_diffusion.openvino/
- gpderetta 4y agoYes, I use [3] and I get 2.4s/iter on my 10 core machine. I was wondering if keras would give additional help here. I'll have to try I guess.