/sdg/ - Stable Diffusion General

Moo my shit up edition

Old bread: >Local install guides
Main: rentry.org/voldy
CPU only: 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
>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

>Other guides
Inpainting/Outpainting: rentry.org/drfar
Upscaling images: rentry.org/sdupscale | rentry.org/b7vcb
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

>Krita Plug-in
flyingdog.de/sd/en

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

>Finetuned Models
huggingface.co/hakurei/waifu-diffusion
huggingface.co/naclbit/trinart_stable_diffusion_v2
Std 1.4 + Waifu 1.2 (linear interpolation "mix"):: mega.nz/file/IwxQTQCb#tsfYZeHrd6U3fg5QCmBazEL-OjOmy1gmhvKAtp1YWFs

>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: frasiercow.png (504x512, 374.13K)

Other urls found in this thread:

hastebin.com/raw/rohosakezi
github.com/MerijnvanEijk/stbl-diffusion-gui-onnx
youtube.com/watch?v=uzQUtbmb6fE
twitter.com/SFWRedditImages

Lewd Diffusion NSFW model, magnet link in the paste at hastebin.com/raw/rohosakezi

Apparently mediocre vs straight WD 1.2 but there if people wanna try it

Spider-Man looks like THIS?

Attached: 02385-3050105176-beautiful gorgeous germanic woman is Spider-Man, full Spider-Man costume, grassy field, focused look, determined face, loving sm.png (512x512, 447.76K)

Attached: 11863-785341838-ornate_intricate_filigree_framed,_vampire_elves_holding_hands_wearing_ornate_intricate_detailed_carved_stained_glass_rose_(((arm.png (1280x1280, 3.41M)

Imagine this technology 5 years from now

Attached: 00029-2465043040.jpg (3456x2304, 1.4M)

i'm excited for the next month, let alone 5 years
gonna be crazy

Anyone know how to abort the render command without killing the entire server?

Attached: 00050-1208348656-Albert Lynch, blond girl, serious, highly detailed, 4k, long hair, elf, paiting, silk dress, flowing dress, blue sky.png (512x640, 467.21K)

Ya got me user, I can not find Waldo.

Ayyymd bros, recently found this, seems like it's not too bad.
github.com/MerijnvanEijk/stbl-diffusion-gui-onnx

Attached: screenshot.jpg (1002x646, 78K)

How is this kind of resolution even possible?!

going to bed now, have fun friends

Attached: 11878-2763311727-bust_shot,_beautiful_female_dark_elf_((long_hair)),_wearing_a_furry_dressing_gown,_ready_for_bed,_(wide_eyes)_(eyeliner),_small_.png (512x704, 348.84K)

I didn't realize it was this bad for you guys

Increase CFG and sampling steps. Might improve results.

god I love this
im not even that big a wondy fan shes just a really good model to test art styles

Attached: unknown.png (478x478, 316.97K)

Cool Gonna try it on my 6900xt how's it run on yours?

Looks *slightly* better than that shitty Stable GRISK on itchio.

Share full prompt before you do?

Attached: 01750-219712568-Shrek_and_Gollum,_The_Creation_of_Adam,_realistic_proportions,_highly_detailed,_smooth,_sharp_focus,_8k,_ray_tracing,_digital_pa.png (512x512, 393.52K)

bust shot, beautiful female dark elf ((medium_hair)), wearing a dressing gown, ready for bed, (wide_eyes) (eyeliner), small_breasts, athletic, muscled,sexy thick_thighs,tight waist , dramatic lighting, by greg rutkowski, Franz Xaver Winterhalter, octane render, ((bed background)), dynamic modeling pose

Negative prompt: ugly, fat, obese, chubby, (((deformed))), [blurry], bad anatomy, disfigured, poorly drawn face, mutation, mutated, (extra_limb), (ugly), (poorly drawn hands), messy drawing, large_breasts, penis, nose, eyes, lips, eyelashes, text,

Steps: 31, Sampler: Euler a, CFG scale: 8, Seed: 1631304899, Size: 512x704

Attached: 11885-1631304899-bust_shot,_beautiful_female_dark_elf_((medium_hair)),_wearing_a_dressing_gown,_ready_for_bed,_(wide_eyes)_(eyeliner),_small_brea.png (512x704, 364.71K)

>vramlet detected

Attached: 00000-2678275702.jpg (3840x1920, 1.96M)

inb4 dudes make fake social media accounts to bait horny dudes into buying them shit and committing fraud in general, getting this tech shut down and/or highly regulated

Attached: newYorkHotDog.png (512x512, 455.48K)

>TFW when it is
Had something similar going with just editing the dml_onnx.py manually, at least it's a bit more civilised now.

Attached: 1639081338701.jpg (588x751, 46.15K)

Thanks user

Attached: 01424-2816152153-Navy,_realistic_proportions,_highly_detailed,_smooth,_sharp_focus,_8k,_ray_tracing,_digital_painting,_concept_art_illustration_b.png (512x512, 301.7K)

If I set it to 150 sampling steps, shouldn't the numbers here be something/150 instead of 113?

Attached: Untitled.png (263x38, 1.71K)

No, I'm not excited. I learned this from video games hype back in 2000.

Attached: 00398-90109-Karl_IV_as_15th_century_crusader,_he_has_very_long_hair,__wears_15th_century_breastplate_armours,_a_cross_on_his_armour,_he_has_.png (512x768, 639.83K)

too white to be a real american

not sure exactly what's happening under the hood but i think for img2img a chunk of it happens two steps at a time so the number in the loading bar is between 50% and 100% of the number of steps you requested

How long does it take to do that outside of VRAM?

youtube.com/watch?v=uzQUtbmb6fE

Attached: apple_2.jpg (1092x720, 188.85K)

Attached: 01410-749656899-Jiangshi,_realistic_proportions,_highly_detailed,_smooth,_sharp_focus,_8k,_ray_tracing,_digital_painting,_concept_art_illustrati.png (512x512, 305.68K)

the more elves the better

Attached: 11911-393239260-bust_shot,_beautiful_female_dark_elf_((long_hair)),_wearing_a_silk_dressing_gown,_ready_for_bed,_(wide_eyes)_(eyeliner),_small_b.png (512x704, 357.65K)

Suuup

Attached: 00000-50_k_lms_2000788304-gfpgan-esrgan4x.jpg (2304x2048, 1.08M)

>old thread still kicking
Faggot OP, as usual.

starting image test

Attached: stable_diffusion_starting_image_test_1.jpg (2540x1080, 778.58K)

Attached: 1663160540192766.jpg (2048x2048, 162.5K)

30-60 seconds per iteration depending on sampler. that one took about 15 mins

Attached: 00022-1344899283.jpg (1856x2560, 897.93K)

How good is SD right now when it comes generating fractals in interesting contexts? I am stuck with MJ right now, and it feels limiting (pic related).

Attached: burningship.png (994x994, 1.6M)