/sdg/ - Stable Diffusion General

Can't sus without the prompt edition

Old Bread: >Local Install Guides
GPU: rentry.org/voldy
CPU: rentry.org/cputard

>AMD Guides
Docker: rentry.org/sdamd
Onnx: rentry.org/ayymd-stable-diffustion-v1_4-guide
Docker on Arch: rentry.org/amddockerarch

>Krita Plug-Ins
github.com/sddebz/stable-diffusion-krita-plugin
flyingdog.de/sd/en/

>Cloud Hosted Installs
Colab: colab.research.google.com/drive/1Iy-xW9t1-OQWhb0hNxueGij8phCyluOh
Paperspace: rentry.org/865dy

>SaaS Sites
txt2img: huggingface.co/spaces/stabilityai/stable-diffusion | beta.dreamstudio.ai | dezgo.com | (NSFW) pornpen.ai
img2img: huggingface.co/spaces/huggingface/diffuse-the-rest

>Other Guides
Inpainting/Outpainting: rentry.org/drfar
SD upscale: rentry.org/sdupscale
Textual inversion: rentry.org/textard
Training: rentry.org/informal-training-guide

>In-Depth Usage Info
Any Forums wiki: wiki.installgentoo.com/wiki/Stable_Diffusion
SD Akashic Records: github.com/Maks-s/sd-akashic

>Tools & Resources
Big index: pharmapsychotic.com/tools.html
Artist styles: rentry.org/artists_sd-v1-4 | urania.ai/top-sd-artists | sgreens.notion.site/sgreens/4ca6f4e229e24da6845b6d49e6b08ae7
Prompt sharing: krea.ai/ | lexica.art/
Dataset: rom1504.github.io/clip-retrieval/ | laion-aesthetic.datasette.io/laion-aesthetic-6pls/images

>Finetuned Models
rentry.org/yrpvv

>Related
/h/ /hdg/: /vg/ /aids/: /mlp/ /ppp/: Official: discord.gg/stablediffusion
NSFW: discord.gg/2ZhBzDGf8u
Training: discord.gg/xVsyrmhQWS

Stability.ai updates: twitter.com/EMostaque

Attached: sus-01.png (512x512, 368.38K)

Other urls found in this thread:

may.2chan.net/b/res/1013613582.htm
intel.com/content/www/us/en/developer/tools/oneapi/overview.html
youtube.com/watch?v=YrlJum5u4uE
anyforums.com/
twitter.com/NSFWRedditImage

if you can add the furry one to the next OP that'd be great
It goes by sdgf

Attached: panda-gen.png (512x832, 493.34K)

may.2chan.net/b/res/1013613582.htm

do I run the install command the same way I ran the prune, with cmd in the stable diffusion folder and "python", space, the code it suggests?
The webui has been running just fine up to now so I'm surprised it says it can't find torch, I though torch was required to get it running in the first place

this one is too easy, but I love how it did it

Attached: 00026-1017705162.png (512x512, 361.8K)

Attached: 9sIlXeE.jpg (3636x5000, 2.49M)

look at those hands on the first try

Attached: 1663208870987-4.png (384x704, 410.96K)

A toast

Attached: a toast.png (512x512, 276.76K)

haha wouldn't it be funny if you pasted the prompt for this haha and then entered the captcha haha, and then clicked post haha

i'm back to generating rainbow waifus

Attached: 00149-1604086923.png (1024x1024, 952.12K)

>Intel Arc A770 is the cheapest 16GB card
>it can't run SD

Is there any hope that will change?

Attached: pepe-desk.png (638x547, 20K)

I got the cmd line version working on my GTX1650 from some youtube video with a quirky british guy ina green morphsuit, but I can't get any version of webui to work, always says it failed to allocate another 20Mb and cuda runs out of memory. The version from the retard voldy guide doesn't see my gpu and fails without a webpage even starting, and the other one I found see's the GPU, opens the webpage, then fails when I tell it to generate even 64*64 images. What do I have to do to make this shit work? Here's a hot starbucks barista from my cmd line version for your attention

Attached: 00004.png (512x512, 397.36K)

Attached: 1663187690378-1.png (1536x1536, 3.22M)

too much porn

Attached: prompt is a secret!!!.png (768x704, 738.14K)

Attached: -911 twin towers burning exploding, by greg rutkowski.png (512x640, 603.99K)

wanna watch that animu

Interesting flowers, can you post the output log?

You can use colab and paperspace.

got some nice honkers

im making suggestive muslim women

Attached: ass2.png (512x512, 261.72K)

lets say i ALREADY have an image and I want to re-create that image in the style of one of the artists in the library

how do i do that? would i use img2img?

also i spent a good portion of yesterday trying to create realistic-looking images, heres a sample

Attached: 00320-4234845880-picture of a teenage redhead, firm breasts, alluring, bokeh, (cleavage), white dress, natural lighting, high detail face, intric.png (512x512, 350.4K)

how go fellow proompters studies in discrete neural networks and automata uwu? I'm trying to solve a multi sum derivation in order to generate a picture of a nice ass, I'm a little lacking in the prompt math so I'll have to do my multivariate calculus reps in order to generate the perfect ass.

Attached: 1663209379677.png (1024x1024, 983.99K)

Attached: st,small,507x507-pad,600x600,f8f8f8.jpg (600x600, 42.82K)

Cute!

problem?

Attached: download - 2022-09-14T233433.631.png (640x640, 392.92K)

Dio Eraclea is really cute.

probably just img2img with a prompt describing the image and then "by " in there

Attached: 00002-4050889244.png (512x512, 402.18K)

Perhaps, in 5-10 years we'll be able to turn such pics(or a collage)+some text into an anime.

Attached: prompt is a secret!!!!.png (768x704, 765.58K)

What are some sci-fi prompts you've been using?

Attached: mfw.png (512x512, 321.88K)

How did you get those hands? Was it a certain key phrase you used in the prompt?

Attached: evil robot girl, wireframe style.jpg (512x512, 127.25K)

holy kino

POST YOUR FUCKING upscaling workflow

SEX?!

>get ai generated image
>push it through waifu2x
>???
>profit

this is borderline subliminal messaging

Please post more of these type of image
I want them all

Attached: E2DC2F98-4D85-4E0F-80A8-E1A300797B08.png (580x545, 452.87K)

if your art isn't coming out right have you tried adding "by Takamichi"?

All these OC characters, but do we have a personification of the AI itself yet?

Attached: 1365440750037.jpg (1024x1980, 910.03K)

I don't know if this one is subtle or obvious.

Attached: girl.png (512x512, 204.91K)

Attached: 1663004067598181.png (512x512, 393.17K)

cringe.

Attached: monkey_lemons.png (512x512, 356.79K)

charred tree stump with black mushrooms, red mold, white mushrooms white mold, grass, branches, tree branches, white leaves, white flower, charcoal, sharp black streaks, nature, 4k, intricate, extreme detail

seed: 3538582301
crf: 19.5
use img2img

>correct looking hands
not ai

Install it in the C drive and don't run webui-user.bat as admin

Make sure you installed python w/ BASH

you need at least 8 gigs of vram from what I heard

>waifu2x

To AI!

Attached: 1635290001835.png (512x512, 336.45K)

>Is there any hope that will change?

If PyTorch gets oneAPI support which is what Intel GPUs will use (as opposed to Nvidia's CUDA), I think that would be great.

intel.com/content/www/us/en/developer/tools/oneapi/overview.html

Attached: Intel-oneAPI-logo-scaled.jpg (2560x2240, 139.26K)

yeah the issue with that is it completely bonks up my image. see attached. it changes the position of the figures and such. anyone know how to simlpy change the art style and medium?

Attached: sample-rotated.png (1289x1792, 1.85M)

This is a very interesting prospect! If only it didn't try to change things...

Attached: file.png (1288x425, 282.16K)

literally nothing
made during the first days of the leak

Attached: explorer_TXUqBpH2e8.png (1760x928, 1.09M)

lower your denoising strength, increase CFG

Turn denoising down and do iterations, feeding the best image back into img2img gradually

Post your prompt music
youtube.com/watch?v=YrlJum5u4uE

Attached: 5yc828.png (512x512, 376.92K)

I mean, I have the local CPU version running (as I don't have a graphics card) and I'm getting pretty decent results

reduce denoising strength

Using this tech without photoshop is like bashing two rocks together to make fire. Making backgrounds and characters separate in SD and then putting them together manually, taking several different renders of the same thing and then creating a composite of the best features of each, and then running it back through img2img. I feel like using it this way I can truly make anything I want and no longer have to worry about the limitations.