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
ProjectAIO_v5
What's new?
Rebuilt from the ground up.
Now includes Cine Diffusion -https://civitai.com/models/50000/cine-diffusion
FAQ
Comments (7)
can you please tell me fp16 is better or fp32 ? also if the file six=ze is bigger that means better?
FP16 and FP32 are floating-point numbers. FP16 uses 16bits, while FP32 uses 32 bits. For the purpose of image generation (using the model) FP32 isn't necessary. Since the model is making predictions based on already learned weights and no significant change to the architecture or weights of the model are made while generating images, using a lower precision data type is recommend. FP16 reduces memory consumption and takes up less space on your storage device while not having any noticeable loss when compared to single-precision (FP32).
~TLDR : For almost all common use cases, FP16 is the "better" version.
Worth mentioning FP32 is usually preferred if you plan to merge it with other models.
Better use fp32 to train a model, but fp16 is a better choice when you just want to generate imgs
Old comment, but in case others are reading this:
If you're using an older gpu, you may run into rendering problems when trying to use fp16 models. It can be remedied with a launch command(I forget which at the moment as I upgraded my gpu and haven't had to use the command for a long time.) If that's the case, you might as well try the fp32 version of a model first, as you won't have the problem with that one(keep in mind though that you will use up more memory using the fp32 version).
If you're using a 30 or 40 series card then always use the fp16 model for rendering. They are the newer cards and are set up to run fp16 perfectly with no problems. You'll also get a speed benefit from doing this and won't use as much memory.
And ofc, as stated previously by others, use fp32 for merging and training(regardless of your card series), and fp16 for rendering.
Doing an in-painting version for 5? or we can still use the 3?
Hello and thank you for the great work. I have a question! I really loved your eris model. Can you tell me how to bring that style in this AIO model with some prompts? (eris is really awesome with high quality Cg-like illustration, especially for skin texture)
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.