Project All In One (AIO)
Please Red Description
Questions/Feedback/Updates
Visit my thread on the Unstable Diffusion Discord
Buy me a coffee ❤
https://ko-fi.com/ndimensional
All donations will be used to fund the creation of new Stable Diffusion fine-tunes and open-source AI tools.
This model is a merge between all of my previously released models and, some unreleased models.
Going forward, all future models and updates to existing models will also be added/updated on this model.
ChangeLog :
12/1/23 V9
Rebuilt from the ground up (as usual).
Included updated models for:
Unreleased version of SciStyle
Added 4 upcoming fine-tunes
7/31/23 V6
Rebuilt from the ground up.
Added models : Cine Diffusion, Electric Eden, Based Model, Another Damn Art Model (ADAM), Lomostyle, Doomer Boomer, and an unreleased Anime fine-tune
Recommended Settings
Clipskip: 1
ENSD: 31337
Sampler:
DPM++ SDE Karras, 30 - 40 steps
DPM++ 2M Karras, 30 - 60 steps
Heun, 20 steps, Sigma Churn = 1
Euler, 20 - 70 steps, Sigma Churn = 1
These parameters are not strictly required, experiment around with other samplers and parameter values. You might find something that works better for you.
Check out my other models (also the models used in the merge)
SDXL
Boomer Art Model - https://civarchive.com/models/163139/boomer-art-model-bam
SD1.5
Doomer Boomer - https://civarchive.com/models/118247?modelVersionId=128239
Lomostyle - https://civarchive.com/models/109923/lomostyle
Another Damn Art Model (ADAM) - https://civarchive.com/models/104898?modelVersionId=112493
Based Model - https://civarchive.com/models/83991?modelVersionId=89262
Electric Eden - https://civarchive.com/models/64355/electric-eden
Cine Diffusion - https://civarchive.com/models/50000/cine-diffusion
WonderMix - https://civarchive.com/models/15666/wondermix
Experience - https://civarchive.com/models/5952/experience
Elegance - https://civarchive.com/models/5564/elegance
VisionGen - Realism Reborn -https://civarchive.com/models/4834/visiongen-realism
LoRA
Pant Pull Down - https://civarchive.com/models/11126/pant-pull-down-lora
Special thanks to rocp for the witch and French maid prompt used in the sample pictures, @nutrition for copious research/testing of various sampling methods/parameters, and to everyone in the Unstable Diffusion Discord community that make doing this a more fun and enjoyable experience.
Description
Initial Release
VAE Recommended
In webui settings, Uncheck: Ignore selected VAE for stable diffusion checkpoints that have their own .vae.pt next to them
Includes
AIO_3-fp16.safetensors
AIO_3.ckpt
FAQ
Comments (23)
Is this the integrated module of the previous model?
this is so fantastic
Thanks!
This model is a combination of all the previous models I released + some unreleased models I'm working on.
@ndimensional awesome. It's still a perfect model, but I hope to see more developments.
@ndimensional kick-ass, your models are great.
@angkuming1004 Oh there will be plenty of new developments 😉
@ndimensional What is the best example of using this inpaint model?
For example, change the background of the model, change the clothes of the model, change the texture of the model.
Are all of the examples above included? Literally [AIO] LOL!!
@angkuming1004 I'm probably not the best person to ask when it comes to inpainting since I hardly use it. (hence the lack of creativity in the inpainting sample pictures lol).
But from the experience I do have with inpainting and how I've seen others use it, most people seem to use it as a way to fix generations that may have one or two things out of place. Others use it to stylize images, add or remove objects, change objects/style/clothing/ect.. And then there's inpainting + ControlNET which I've never experimented with so I can't say much on that front.
If I find some resources or guides I'll link them to you.
Maybe dumb to ask, but do you think a prompt set will have the same effectiveness within this checkpoint versus using the other checkpoints individually? Thanks
Also, do you have any concern with [Experience] being infused into this? It being non-real/art focused, any fear it will get in the way of when you're trying to compose real-based images.
Thanks for reading
From my testing, the generated output will be somewhere in the same ballpark as the individual checkpoint but won't be exactly the same.
As for your concern, getting all the models to balance nicely; not having a bias towards one particular model, was part of the fun (for someone like me) when making this mix. AIO can do photorealistic images like "Clarity" and "Refined" as well as 3D renders like that found in "Experience". Granted, it's not the most photorealistic model I've made. But it's capable of photorealism.
If you want a pure photorealistic model, my next model is the most photorealistic models I've made. It's trained on 35mm film stills, 24mm APS film, and a database of photographs/portraits that have been processed in Luminar Neo for enhanced skin texturing and lighting. No ETA on when that'll be released but it shouldn't be that far out.
I see, guess it's 'jump around test different checkpoints for your prompt' situation as always. No harm in keeping them all.
Side note, I've found that Experience and VisionGen easily achieve specific things, whereas Elegance and Clarity their usage scenarios and advantages are not quite clear.
(Maybe just I don't know how to use them) (Never tried REFINED)
Anyways looking forward to the cool new model.
I find your models to be some of the best out there. Also you seem to put a lot of effort and research into them. I'm surprised you didn't add Offset Noise yet, which is a much needed fix for SD models.
No hurry but many of us can't go back to pre- Offset Noise world once you enter, have fun testing it out, and thanks.
I'm aware of offset noise, but I appreciate the comment and recommendation!
Yes, offset noise can dramatically enhance the fidelity of generations via lighting/contrast. I'm roughly 60 hours into training a large model that uses offset noise. I'm hoping to get it released by the end of the month! 🤞
I'd love a checkpoint mix of just elegance, and clarity. To be honest, I prefer experience as its own separate thing for digital illustrations type art, rather than photos. I know how to do checkpoint merges myself, problem is I suck at getting a good balance between them. Other than that you probably have the best models on the site for photo realism, other than realistic vision, but yours are my new go to because of the drama surrounding them.
Thanks for the kind words!
If you need any help with merging, you can DM me on discord (links to my discord thread is at the bottom of the model description).
@ndimensional that's, ok this model has actually been been good for my needs, I just have to specify it's a photo with prompts. It does some great sci-fi. For more traditional vintage photos I probably will do a 50/50 blend of clarity and elegance.
man I should say you have the best merges and very good taste for prompts
Did you upload v4.5? I got a notification that a new version of this model was uploaded, but I don't see any links to it in the download section (just AIO v3 is there)
Updates on civitai are currently broken. Model couldn't be uploaded. For some reason a blank model version (no files or sample images) was created - so I removed it.
@ndimensional I see. I remember seeing there were issues with model uploading, but I wasn't sure if this notification was posted before or after Civitai resolved this issue (or if it's been resolved yet)
Still eager to try the newest version, whenever you find time to upload it again
@Kodama Thanks!
Hopefully I can get a confirmation soon on whether or not the issue has been fixed, attempting it again would just push another notification. 🤞
Something happened with version 4.5? I have 5 update notifications.
The update feature on civitai is currently broken. Model couldn't be uploaded.
@ndimensional
I think they have already solved the problem, I already correctly receive updates from other models and loras
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.