Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Try Stable Diffusion's Img2Img Mode

huggingface.co

Post reply on HN
Print view View original

51–60 of 166 posts

Re: Try Stable Diffusion's Img2Img Mode

#51

Post by avocado2 »

web demo for stable diffusion (txt2img): https://huggingface.co/spaces/stabilityai/stable-diffusion

github with gui: https://github.com/hlky/stable-diffusion

dev repo (more features, may have bugs): https://github.com/hlky/stable-diffusion-webui

repo with docker: https://github.com/AbdBarho/stable-diffusion-webui-docker

colab repo (new): https://github.com/altryne/sd-webui-colab

can also run it in colab (includes img2img): https://colab.research.google.com/drive/1NfgqublyT_MWtR5Csmr...

demo made with gradio: https://github.com/gradio-app/gradio

Re: Try Stable Diffusion's Img2Img Mode

#52

Post by lkois »

jtolmar wrote: post #50
comex wrote: post #13

If you have a GPU with >4GB of VRAM and you want to run this locally, here's a fork of the Stable Diffusion repo with a convenient web UI: https://github.com/hlky/stable-diffusion It supports both txt2img and img2img. (Not affiliated.) Edit: Incidentally, I tried running it on a CPU. It is possible, but it took 3 minutes instead of 10 seconds to produce an image. It also required me to hack up the script in a really…

I'm impressed that running it on CPU only made it ~20x slower. How did you do it?

Nah that's normal. It's why GPUs are the usual thing for AI. Any crap, old, weak gpu with 4gb memory would run circles around a cpu

It's often easier to actually get models to run on CPU, due to simpler install configs and more available memory. Just painful to get a result out of it. Which might help keep the install simple, because it's not even worth optimizing

Re: Try Stable Diffusion's Img2Img Mode

#53

Post by altryne1 »

avocado2 wrote: post #51

web demo for stable diffusion (txt2img): https://huggingface.co/spaces/stabilityai/stable-diffusion github with gui: https://github.com/hlky/stable-diffusion dev repo (more features, may have bugs): https://github.com/hlky/stable-diffusion-webui repo with docker: https://github.com/AbdBarho/stable-diffusion-webui-docker colab repo (new): https://github.com/altryne/sd-webui-colab can also run it in colab (includes img…

I'm one of the maintainers (in charge of the UI) for hlky webui repo! And we just updated our own colab, you can find it here : https://github.com/altryne/sd-webui-colab

Re: Try Stable Diffusion's Img2Img Mode

#54

Post by desindol »

hakuseki wrote: post #37

I'm not quite grasping how to use this. I tried uploading a photograph and erasing part of it. But instead of painting in the erased portion, it left the erased area blank and replaced my photograph with an entirely new image.

What you tried is called inpainting. Img2Img needs a rough sketch.

Re: Try Stable Diffusion's Img2Img Mode

#55

Post by asicsp »

comex wrote: post #13

If you have a GPU with >4GB of VRAM and you want to run this locally, here's a fork of the Stable Diffusion repo with a convenient web UI: https://github.com/hlky/stable-diffusion It supports both txt2img and img2img. (Not affiliated.) Edit: Incidentally, I tried running it on a CPU. It is possible, but it took 3 minutes instead of 10 seconds to produce an image. It also required me to hack up the script in a really…

Can you share your script for running on CPU?

Re: Try Stable Diffusion's Img2Img Mode

#56

Post by MichaelDickens »

comex wrote: post #13

If you have a GPU with >4GB of VRAM and you want to run this locally, here's a fork of the Stable Diffusion repo with a convenient web UI: https://github.com/hlky/stable-diffusion It supports both txt2img and img2img. (Not affiliated.) Edit: Incidentally, I tried running it on a CPU. It is possible, but it took 3 minutes instead of 10 seconds to produce an image. It also required me to hack up the script in a really…

What kind of GPU do you have? It takes several minutes to produce an image on my 1070.

Re: Try Stable Diffusion's Img2Img Mode

#58

Post by kgwgk »

comex wrote: post #13

If you have a GPU with >4GB of VRAM and you want to run this locally, here's a fork of the Stable Diffusion repo with a convenient web UI: https://github.com/hlky/stable-diffusion It supports both txt2img and img2img. (Not affiliated.) Edit: Incidentally, I tried running it on a CPU. It is possible, but it took 3 minutes instead of 10 seconds to produce an image. It also required me to hack up the script in a really…

There are as well forks for the GPU in Apple’s M1 chips:

https://github.com/magnusviri/stable-diffusion

Re: Try Stable Diffusion's Img2Img Mode

#59

Post by desindol »

rvz wrote: post #48
keepquestioning wrote: post #6

Can't help but see this as a friendly herald for a dystopian future.

There you go. We were promised quite far fetched things like 'flying cars', 'time travel', 'life extension' and 'universal basic income'. Instead we get Deep Learning AI's being used for generating and faking sentences, images, videos, voices, code and digital art all being trained on mountains of data in data centers all significantly contributing to the already burning up of the planet to no benefit and no efficien…

You must be kidding it’s hard labor creating images and you are complaining that it’s going to be easier in the future? You are complaining that the skill gap is eradicated and it now only depends on your ideas to produce compelling work? That’s absolutely insane everything that takes labor intensive tasks and makes them basically free is the future. This is not dystopian it’s our future, a future in which you and your capabilities don’t matter. It’s exactly the future we need and doing it opensource is exactly the way to keep it out of harms way of turbo capitalism. You are missing the point so hard it’s not even funny. Flying cars for the 1% is dystopian.

Re: Try Stable Diffusion's Img2Img Mode

#60

Post by stingraycharles »

MichaelDickens wrote: post #56
comex wrote: post #13

If you have a GPU with >4GB of VRAM and you want to run this locally, here's a fork of the Stable Diffusion repo with a convenient web UI: https://github.com/hlky/stable-diffusion It supports both txt2img and img2img. (Not affiliated.) Edit: Incidentally, I tried running it on a CPU. It is possible, but it took 3 minutes instead of 10 seconds to produce an image. It also required me to hack up the script in a really…

What kind of GPU do you have? It takes several minutes to produce an image on my 1070.

A 1070 isn’t very powerful for ML compared to more recent GPUs, so several minutes sounds about right.
Post reply on HN