Stable Diffusion, NovelAI, Machine Learning Art - AI art generation discussion and image dump

  • 🐕 I am attempting to get the site runnning as fast as possible. If you are experiencing slow page load times, please report it.
1665959523169355.png


3516988637-(extremely detailed CG unity 8k wallpaper), masterpiece, quality, (((full body))), ...jpg
2801489177-(extremely detailed CG unity 8k wallpaper), masterpiece, quality, (((full body))), ...png
3483920095-(extremely detailed CG unity 8k wallpaper), masterpiece, quality, (((full body))), ...png
2367315031-(extremely detailed CG unity 8k wallpaper), masterpiece, quality, (((full body))), ...png
2367315033-(extremely detailed CG unity 8k wallpaper), masterpiece, quality, (((full body))), ...png

1308407524-(extremely detailed CG unity 8k wallpaper), a (cute girl), petite, (((upper body)))...png
2645696763-(extremely detailed CG unity 8k wallpaper), a (cute girl), petite, (((portrait))), ...png
tmp53y9pe8y.png
tmpzhitxc3f.png
Ew.png
 
Last edited:
So I've been meaning to ask, are there any good sites to generate images without the whole "SUB NOW FOR 100 JEW TOKENS"?
Dezgo has been the only one that's legit and can generate anything for free, but only does a single image at a time.


My current PC can't handle doing everything remotely so if anyone's got a list of good sites, especially for the anime side of things, send them my way 👍
 
  • Thunk-Provoking
Reactions: IAmNotAlpharius
I'm sorry if this has been answered before but as someone with an AMD GPU is it possible to set this up with webui on Windows or even Linux ? Last time I tried was a complete disaster and nothing worked, and I'd really like to be able to use this.
 
I'm sorry if this has been answered before but as someone with an AMD GPU is it possible to set this up with webui on Windows or even Linux ? Last time I tried was a complete disaster and nothing worked, and I'd really like to be able to use this.
Using Linux, in theory yes. As I don't use either that system or hardware, I can't confirm how well it works (or even if the guide has been updated over time) but there's a guide for AMD Windows and Linux, with the Linux guide using the Automatic1111 web UI.
 
I see its finally learning how to do hands.
Using Linux, in theory yes. As I don't use either that system or hardware, I can't confirm how well it works (or even if the guide has been updated over time) but there's a guide for AMD Windows and Linux, with the Linux guide using the Automatic1111 web UI.
What about intel ARC?
 
So I've been meaning to ask, are there any good sites to generate images without the whole "SUB NOW FOR 100 JEW TOKENS"?
Dezgo has been the only one that's legit and can generate anything for free, but only does a single image at a time.


My current PC can't handle doing everything remotely so if anyone's got a list of good sites, especially for the anime side of things, send them my way 👍
As long as you've got new email addresses you can keep using midjourney. It generates on their server so it's just using discord. It gives you 20 free generations per account.

Alternatively you could pay and keep generating without making new accounts.

Except for nsfw or gore, midjourney is one of the best and certainly one of the easiest to generate with and it was trained on beautiful art.
 
more cool stuff, I love experimenting with sketchy stuff that looks like someone actually drew it! also thank you all for all the likes on my posts ur all cool and you guys generate cool shit!
 

Attachments

  • 00319-2266749948-Ken Sugimori, 1girl, short spiky black hair, brown eyes, ((wearing school uni...png
    00319-2266749948-Ken Sugimori, 1girl, short spiky black hair, brown eyes, ((wearing school uni...png
    974.2 KB · Views: 46
  • 01175-2730549552-((official art)), 1girl, short spiky brown hair, seifuku, face, smile, head-s...png
    01175-2730549552-((official art)), 1girl, short spiky brown hair, seifuku, face, smile, head-s...png
    749 KB · Views: 41
  • 01260-2625718977-((official art)), geralt of rivia, ((mature)), wearing heavy dark fantasy arm...png
    01260-2625718977-((official art)), geralt of rivia, ((mature)), wearing heavy dark fantasy arm...png
    476.6 KB · Views: 39
  • 01192-1723076471-((official art)), 1girl, ((mature)), flowing blonde hair, wearing a old-Europ...png
    01192-1723076471-((official art)), 1girl, ((mature)), flowing blonde hair, wearing a old-Europ...png
    841.3 KB · Views: 44
  • 00910-208811921-Ken Sugimori, 1girl, nude, short black hair, cyberpunk aesthetic, (((cyberneti...png
    00910-208811921-Ken Sugimori, 1girl, nude, short black hair, cyberpunk aesthetic, (((cyberneti...png
    980.3 KB · Views: 46
  • 00326-3566842415-Ken Sugimori, 1girl, short spiky black hair, brown eyes, ((wearing tomboy sch...png
    00326-3566842415-Ken Sugimori, 1girl, short spiky black hair, brown eyes, ((wearing tomboy sch...png
    820.9 KB · Views: 45
So I've been meaning to ask, are there any good sites to generate images without the whole "SUB NOW FOR 100 JEW TOKENS"?
Dezgo has been the only one that's legit and can generate anything for free, but only does a single image at a time.


My current PC can't handle doing everything remotely so if anyone's got a list of good sites, especially for the anime side of things, send them my way 👍
Just use a google colab notebook or whatever unless they've cracked down on this, but last I heard still work. Search the S-D rentry pages and you'll find some already set up, with frontends even I'm pretty sure.

There's a few platforms like this, they're for machine learning education mainly and give you decent performance for free with the only limitations being kicking you off if you're inactive for too long or using too much resources for hours at a time. It won't be super fast but I've used colab and kaggle or something for crunching text transformer stuff in the past and IIRC they were a little better than my old laptop I tried S-D on, so probably 20 second gens at lower res and you'll be able to just have batches running in a tab while you do other shit.
 
  • Informative
Reactions: A Hot Potato
I would like to thank @Starved Artist for mentioning Civitai, which I was previously unaware of. So many models, so little time...

By far, the largest benefit to how the Hires Fix in the web UI works now is time saved; it's so much easier to generate high quality, high resolution images at a much faster pace than previously. Naturally, I used this to play around with a new model. It can also do semi-decent hands fairly often, but I was more interested in faces.
Test image.jpg00054 - Lake.jpg00056 - Pose.jpg
00070 - Portrait.jpg00114 - Scifi.jpg
00156 - Fox.jpg00154 - Winter Queen.jpg00290.jpg0029.jpgFoxFace.jpg
A little bit of photoshop fix up was needed on some of these.
Model could handle more facial variety than what I posted, but I'm only posting what I chose to upscale.
Because Year of the Wabbit and whatnot.
00071.jpgbunnyEyeFix.jpg
Different model; just needed some inpainting and photoshop to fix things.
00004.jpg
Because I felt like it was contractually obligated.
0035.jpg
And lastly, because it amused me.
00125 - lol.jpg
That's not how perspective works.
Much more to test and experiment with but I already think DreamShaper 3.3 is a keeper.
 
Would an nvidia quadro work out running on Linux for the sake of stable diffusion?
 
Back