The result of me deciding to put on the merging wizard hat on once more, much like with my Wild Ponies models. Mostly for my own edification, with many, many failures along the way, this finally strikes the balance of increased detail and nice aesthetics - or so I hope, anyway!

V1.1

• Trained up to Illustrious 1.1 (will eventually redo from scratch using it as a base)

• High Resolution Support

• Better Natural Language Prompting

• Much less shimmering artifacts!

Comparison with some models I like and Base IL 1.1:

So, some good results, but stuff like the Slime Girl clearly still needs some work!

Alpha

Base Model: Illustrious 1.0 - Supports higher resolutions and preferred for generation!

Also supports some level of simple NLP Prompting! (see several examples here)

Note: Some LoRAs may need some extra strength to them, as there is some XL models in the recipe, but most work fine.

Otherwise, just use with same samplers, CFG, as you do with your other IL models, it should work nicely, but steps benefit from being a little higher.

This is a first viable release; I may experiment with training and merging in LoRAs to further shift the base style to the one I originally sought to achieve but fell short of.

Feedback, and advice on merging welcome! Still very much a novice looking to get better!.