What is it?
Try Stable Diffusion's Img2Img Mode
11–20 of 166 posts
Re: Try Stable Diffusion's Img2Img Mode
#12What are your best prompts so far?
"john wayne as captain america" is spot on. It has so many images of both to work with.
Re: Try Stable Diffusion's Img2Img Mode
#13If 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 gross way. Perhaps there is a script somewhere that properly supports this.
Re: Try Stable Diffusion's Img2Img Mode
#14Is there a tutorial on how to run this on Paperspace?
Re: Try Stable Diffusion's Img2Img Mode
#15Re: Try Stable Diffusion's Img2Img Mode
#16Well. Apparently "penis on park bench" is an "inappropriate" query. However, "lot of child dying in a massive fire" is an a-okay query with some "interesting" results. Prudes are such a weird bunch.
I think of this like a "first post" filter, to cut down on the noise. Drawing dicks on things is something a lot of kids do. You want the trolls to have to work a little.
Re: Try Stable Diffusion's Img2Img Mode
#17Re: Try Stable Diffusion's Img2Img Mode
#18You can also try it out on Replicate: https://replicate.com/stability-ai/stable-diffusion
Re: Try Stable Diffusion's Img2Img Mode
#19Looks pretty hard to get it to run. Queue full every time.
You can also run it on Replicate without a queue: https://replicate.com/stability-ai/stable-diffusion (set “init_image”)
Re: Try Stable Diffusion's Img2Img Mode
#20Well. Apparently "penis on park bench" is an "inappropriate" query. However, "lot of child dying in a massive fire" is an a-okay query with some "interesting" results. Prudes are such a weird bunch.
I’m not gonna lie, I’m quite disappointed at all the anti-NSFW shenanigans.
I'm not gonna lie, I'm quite disappointed by the wilful obstinance on HN whenever there is a discussion of anti-NSFW or anti-racism filters that are present on today's artificial intelligence research.
First, the filters are only on the interface. The research is public. That's why there has been an explosion of new implementations. You are welcome to run the code yourself and make the horniest model you like.
But more importantly, this comes up every time, and everyone acts SO confused. But why? Artificial Intelligence is MASSIVE mainstream news. New advancements are arriving daily.
Do you want the news stories to be dominated by whatever heinous shit some bad-faith giggly teenager with a call of the void to offend as much as possible (i know, i was one) is able to generate?
Do you want the discourse to be dominated with "Won't someone think of the children?" blocking legitimate research and progress?
Europeans always like to dunk on Americans for being "prudes", but only a tiny section of Western Europe is progressive enough to not mind random nipples on their bus ads and television. All of Eastern Europe, Africa, India, and China are culturally still fairly conservative about sex - at least out in the open.
People think that preventing racist or sexual use of these models is indulging the prudish mores of America. But that itself is a very narrow perspective that ignores the perspectives of billions of people in the world.