Aesthetic Quality Modifiers - Complete
Training data is a subset of all my manally rated datasets with the quality/aesthetic modifiers, including only the masterpiece, best quality , or very aesthetic tagged images (7273 total).
Superset and final of models in the Aesthetic Quality Modifiers Collection.
ℹ️ LoRA work best when applied to the base models on which they are trained. Please read the About This Version on the appropriate base models and workflow/training information.
Recommended prompt structure:
Positive prompt (quality tags at the end of prompt):
{{tags}}
masterpiece, best quality, very aestheticGeneration Settings:
Previews are generated in Forge with upscaling and adetailer.
For Noobai V-Pred, a ComfyUI workflow reference with DynamicThresholding, Upscaling, and FaceDetailer can be found here: https://civarchive.com/posts/11457095
Description
Trained on Illustrious-XL
FAQ
Comments (10)
Does this replace the H variant also?
Yes, it is a superset of that dataset
What is the difference between all these Quality Modifiers loras?
Some experimenting to see the result of manually rating a curated >22,000 image dataset.
Separated datasets for masterpiece (266), best quality (4882), and this combined masterpiece+best quality+very aesthetic (5332) LoRA.
For quality, evaluated anatomy, perspective, & depth.
For aesthetic, evaluated composition, backgrounds, light, & color.
@motimalu Which one would you consider most "polished" or well-rounded lora for quality of the images?
@ComradePedro If you mean a lighter effect as an improvement to any generation and not any concepts in particular that I've trained other LoRA for, I'd recommend the relatively much smaller (~300 image) Masterpieces LoRA at the moment.
(Answered a similar question here)
Can I use this lora on Noobai v pred model？
One before/after post using this LoRA's e pred version on the v pred model here:
https://civitai.com/posts/11080293
As far as I can tell and from what others have said, LoRA's trained on Noobai v pred or e pred can be used interchangably.
I am still testing v pred a bit and updating my local workflows for it, so can't really advise on the best use of it though.
Mentioning that these Loras can't be used with the on-site gen of noobai, because noobai is set as a SDXL 1.0 base model and these Loras, Illustrious. They simply don't show up on the generator for Noobai.
Thanks @Churrus - I matched the base model as SDXL for the noobai LoRA's, so they should appear as an option
Details
Files
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.