/sdg/ - Stable Diffusion General

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

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

>M1 Supported Fork
github.com/lstein/stable-diffusion

>Cloud Hosted Install
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
inpainting: huggingface.co/spaces/fffiloni/stable-diffusion-inpainting | inpainter.vercel.app/paint

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

>Other Guides
Inpainting/Outpainting: rentry.org/drfar
Upscaling images: 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: lexica.art/ | krea.ai/
Dataset: rom1504.github.io/clip-retrieval/ | laion-aesthetic.datasette.io/laion-aesthetic-6pls/images

>Finetuned Models
rentry.org/yrpvv

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

Stability.ai updates: twitter.com/EMostaque

Attached: imsdf.png (512x512, 732.94K)

Other urls found in this thread:

rentry.org/sdamd
youtube.com/watch?v=YAn8FVUp090
github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Features
twitter.com/SFWRedditGifs

I want a big breast wife ;_;

responding to a post from yesterday before I went to sleep

so a few things the guide does not do a good job explaining
1. it doesnt tell you that youll need CUDA installed till after you get it running in the first place and you need CUDA to run other shit
2. once you get it THEN you go to pytorch, where it says local install you select pip and whatever CUDA you installed, copy that into a normal command prompt
3. once you have torch you can now do things like prune the waifu diffusion model and like I was doing merge 2 models

I forgot to save it before moving on so enjoy a crop of charm

Attached: file.png (643x639, 767.67K)

>Base SD model understands latex perfectly
Thank you Emad, I love the machine

Attached: 00101-4081906072-painting of two people wearing ((((latex clothing)))) sitting in an ornate throne-room, beautiful faces, castle backdrop, vivid .png (1024x1024, 1.17M)

another beauty

anyone got anymore

and what promps do u use

Does Ilya Kuvshinov even have a reason to exist anymore? AI can do her job in second now.

I want a flat chested wife ;u;

>just signed up for a bunch of drawing courses last month
How good do I need to be able to draw for my pics being useful in creating unique prompts?

Attached: 1621562897080.png (500x500, 384.36K)

How long until someone creates a .exe making things actually acessible?

It really can't yet.

My prompts are too powerful for you, promptlet. You best find a prompt seller with weaker prompts.

Attached: 00004-432827020.png (512x512, 245.01K)

Will generating pics nuke my laptops GPU? It's a 2080MQ

Attached: 1601363568258.jpg (640x847, 42.77K)

>Docker: rentry.org/sdamd
WSL2 does not support AMD GPU.

His. Ilya is a male name. Male surnames end in -ov. Female ones in -ova.

>IndexError: index 1000 is out of bounds for dimension 0 with size 1000
what's this about then?

LET GOOOOOO

Attached: Screenshot_20220915-164724_Discord.jpg (1080x2400, 785.77K)

The future. Is now.

Attached: 1663273230670193.webm (448x832, 2.98M)

technology was a mistake

Attached: unabomber.jpg (274x253, 45.39K)

I'm stupid. explain further. is this something that can be plugged right in and I'll see the upgrade in a few hours/tomorrow or is this more of a "you'll see something in a few days"?

That's some decadent cyberpunk clothing style

That's like over a week old bruv

Fuck, i always imagined him as a cute russian girl, i feel cucked now

now make her pregnant

asking unironically, what are some good uses for this outside of coom? I don't mean uses in future or how it could be applied later, I mean currently, what's the draw for solo usage?

Attached: 00167.png (1024x1536, 1.57M)

Hahahah look at him and laugh, get fucked.

[spoiler]I thought the same[/spoiler]

Selling NFTs

its having some trouble making my wife from just a text prompt
gonna need to find one good base and then refine it when I get home

Attached: 1663151790418678.jpg (690x538, 154.38K)

Pen & paper RPG campaigns

Attached: leonhard2.png (512x512, 313.89K)

You know how artists use references, right?
This is the ultimate reference / composition / idea generator.

Beyond that, it helps to expand upon basic ideas and structures with the inpainting / outpainting (fixing errors, reducing time spent, etc).

I've been using it to create variations of my own designs/concept art via img2img. Though I won't use the raw output in any actual projects, its already given me new ideas and helped me conceptualize design routes I hadn't thought of myself. It's like showing work to my peers and having them give me feedback on possible alternative designs. I really like it.

>Stable diffusion. Create one Deanna Troi, (Futanari), 10 Inch Penis(Uncut), dominating Reginald Barclay, Cum, Sissy, Unreal Engine, soft lighting

Attached: barclay.jpg (680x391, 49.44K)

We were promised a lewd tune.
They said it'd take a few days.
That was a week ago.
WHERE IS IT.

Creating media we actually want to consume instead of having malicious corporations dictate what we must like.

Right now, in the current state, drawfags use is as a reference tool, as a backdrop photobashing tool (something that consumes a lot of time), as a style transfer tool (it's unstable in that though). Studios have resources to spare and can train it on their own material to severely cut pixel pushing time.

Test

i fucking love this booba generator

Attached: booba.png (564x528, 444.47K)

Why is this shit so hard to install? Why can't they just give us an exe file to click and install like normal programs do? Are they trying to filter out less tech savvy people?

Thick shortstacks

Attached: 3555.png (512x768, 533.88K)

in the OP

img2img batch is broken for me now, have to restart entire thing each time i want to interrupt...

I'm sure you don't actually use the term shortstack to get that body type... do you?

also testing

Plenty of goyslop out there to consume for braindead normie types. If they can't rub two brain cells together to produce spark enough to install two or three dependencies, then run a fucking .bat file, I don't know what to tell you.

Halfling
I'm going to move to goblin girls next

It has to do with how the data is coded.
Yes, you could package it as an exe file and run it as a stand-alone program, but with that method comes more memory use and less efficient processing.

The current method gives a simple install (seriously, it isn't hard if you READ) and retains flexibility for configuring it / customizing it to work on different systems with different needs.

Because you need a bunch of nvidia libraries and shit, which aren't easily portable and there aren't any good end-user solutions that work with an interpreted python environment

what does this mean better images and less creepy faces?

yes, they are trying to filter faggots who need spoonfeeding and will shit up the thread, like (You) in particular

it's literally one user slapping together a python abomination, on top of some crude code not intended to be used as a final product

Stop ban evading and take it like a man.

Oh my...

Attached: mia.jpg (2944x5012, 2.94M)

use this and follow the instructions. its very easy to install
github com/AUTOMATIC1111/stable-diffusion-webui

It's actually really really bad at general composition.

Midjourney is better at that, but all AI ends up being too samey in that regard so as far as illustration goes, it's not too amazing in that regard.

For ideas it's also not too good either. You might end up getting caught just spending hours generating random images for shit you might have been able to do on your own. Name of the game is time saved so it falls a bit short in that regard too.

One strong point non-artist seem to miss is color/values. AI generally doesn't fuck that up, and particularly Midjourney has some good color work. Raw SD can tend tobe a bit washed or have a palette that's a bit all over the place, but it's still pretty functional as a crutch if values are your weakness, and it is the weakness of most low tier artists.

You can generate something, color pick from it and basically do away with some of the guesswork.

youtube.com/watch?v=YAn8FVUp090

Attached: trumpup2ps.png (1792x3072, 3.65M)

me too fren

Attached: 000000.jpg (2304x2304, 1.4M)

Quick! Reply with your best generations!

Attached: dgsfgswdf.png (512x512, 423.31K)

What’s the point of (((brackets)))

given how amazing this timeline is...
how long till we have actual holodecks?

Attached: ai.jpg (1469x660, 89.66K)

RTFM
github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Features

I'm having a problem posting a funny futanari attempt in /h/, I had one warning for nsfw here very early on and I have been a good boi since then

>The current method gives a simple install (seriously, it isn't hard if you READ) and retains flexibility for configuring it / customizing it to work on different systems with different needs.
There's a couple of main issues that I have had installing this. Firstly, my laptop doesn't have a gpu so i can't use the normal install method. Secondly, the large number of steps creates lots of opportunities to fumble the install and makes it hard to figurr out where exactly you messed up.

Give a keyword more weight