Stable Diffusion Link Dump

[Freak]

Hey, I've been playing around with Stable Diffusion lately, and I figured I'd just dump a bunch of links for how people can get started and all the different ways to use it.

Common fork that lets you use the model in a browser UI:
https://github.com/AUTOMATIC1111/stable-diffusion-webui/releases

Weights for the 2.1 version:
https://huggingface.co/stabilityai/stable-diffusion-2-1

 

Amazing site with thousands of different fine tuned models that you can download:
https://civitai.com/

 

This is the Github page for Stable Riffusion. A model that creates music by generating audio spectrogram images:
https://github.com/riffusion/riffusion-app

Riffusion weights:
https://huggingface.co/riffusion/riffusion-model-v1

[Freak]

If anyone has any specific fine tuned models that they really like, you should post them here! Here are some that I like:

DreamShaper
https://civitai.com/models/4384/dreamshaper

Dungeons and Diffusion
https://civitai.com/models/90/dungeons-and-diffusion-v3

 

[cwade12c]

I am a pretty big fan of Elldreth's models. Vivid, Retro, Lucid, and Dream.

Here's some examples of Vivid

https://civitai.com/models/2747/elldreths-vivid-mix

Positive	Negative	Output
award winning waist up photo of a beautiful fantasy warrior, wearing torn black leather armor, tarnished and scratched silver pauldrons, worn leather gloves, short blonde hair, bright blue eyes, dense forest in background, dark, ominous, foggy, high contrast, shiny skin, soft lighting, backlighting, bloom, light sparkles, chromatic aberration, sharp focus	3d, cartoon, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, artist name
award winning waist up photo of a rugged dark science fiction space marine, wearing scratched and dented space marine gear, middle-aged, short copper hair and beard, green eyes, inside spaceship cockpit, electronic circuits hanging loose in background, flashing control panel in foreground, science fiction, action, dark, tools on floor in background, dark red lighting, high contrast, shiny skin, backlighting, bloom, light sparkles, chromatic aberration, sharp focus	nsfw, 3d, cartoon, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, artist name
award winning waist up photo of a beautiful sexy action hero, wearing dark green worn tank top, torn and dirty khaki cargo pants, indiana jones hat, leather whip, long blonde hair, braided, green eyes, swinging over river, in the jungle, dense jungle in background, sweaty, face flushed, large jungle bush in foreground, cave in background, action, sunny, jungle plants in background, high contrast, shiny skin, backlighting, bloom, light sparkles, chromatic aberration, sharp focus	nsfw, 3d, cartoon, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, artist name

[cwade12c]

Some other information and links I have come across....

I recommend launching the SD Web UI (automatic1111) with the following argument for better performance:

--xformers

This site has a lot of tutorials about SD, from technical explanation to how to do different things:

https://stable-diffusion-art.com/

I also have been having good success with textual inversion, hypernetworks, and LoRAs. These three are also much easier to make yourself on a budget, compared to checkpoints.

For text inversion, check out theally's releases.

https://civitai.com/models/3485/princess-style

https://civitai.com/models/2289/psycho-style

https://civitai.com/models/1998/autumn-style

https://civitai.com/models/2032/empire-style

 

I think my favorite hypernet right now for interesting results is Luisa's Magiclight.

https://civitai.com/models/4077/luisap-magiclight

 

As cool as LoRAs are, I don't have any favorites to recommend but I do suggest checking out rentry's training guide

https://rentry.org/lora_train

 

p.s. don't get pickled! Always use SAFETENSORS for security reasons.

[SoArtificial]

I also became quite obsessed with stable diffusion. So obsessed I decided to make my own website. I create articles on news and guides all about Artificial intelligence.

My site is SoArtificial 

I have created a few guides about stable diffusion one of my more popular ones is my stable diffusion cheat sheet but I also do guides on how to create realistic people and how to install stable diffusion on a windows machine. The more I learn the more I add to the website. I hope some of the articles found on my site do help people out. 

[George Regan]

I'm intrigued by the concept of Stable Diffusion. Managed to implement it in some of my AI models, and saw noticeable improvement. Also, I sometimes use a diffusion tool, 0xAi, to generate stable image hints using diffusion. If you are interested in this, you can also try