Aesthetic Quality Modifiers - Absurdres

(Experimental) This LoRA is part of an attempt to fine-tune a Pony Diffusion derived model to not require the use of score quality modifiers, with the hopes of also gaining more compatibility with other SDXL anime model merges.

Training data includes a selection of only top-scoring (>75%) absurdres tagged booru images.

Manually (human) tagged in the same format of quality modifiers as the Animagine XL V3.1 checkpoint.

These modifiers can now be applied to the end of the prompt

Quality Modifiers

Quality Modifier	Score Criterion
masterpiece	        > 95%
best quality	        > 85% & ≤ 95%
great quality	        > 75% & ≤ 85%
good quality	        > 50% & ≤ 75%
normal quality	        > 25% & ≤ 50%
low quality	        > 10% & ≤ 25%
worst quality	        ≤ 10%

Aesthetic Tags

Aesthetic Tag	       Score Range
very aesthetic	       > 0.71
aesthetic	       > 0.45 & < 0.71
displeasing	       > 0.27 & < 0.45
very displeasing       ≤ 0.27

Prompt structure

Positive prompt (no score tags required, quality modifiers at end of prompt):

{{tags}}
masterpiece, best quality, very aesthetic, absurdres

Negative prompt:

low quality, worst quality, 3d, monochrome, (censor), render, text, watermark