bimbostyleThreeU (DeliberateV2) - LORA

This is model uses the exact same input as bimbostyleThree but was trained on deliberate_V2 which hopefully means that it's better applicable to other .ckpt in comparison to the normal "Three"

Since i mainly use "draw style" ckpt i can't really say anything about the full potential of this Lora on other ckpt.

Combining this Lora together with One/Two/Three in a "draw style" ckpt give interesting results show in the preview for example in pic 3 with lyriel and i can only recommend trying it out. (i don't know if the same applies to "real" ones as well)

If you mix my models, i would recommend to always put the stronger weighted Lora in the front. i.e.

lora:bimbostyleThree:0.5:lora:bimbostyleTwo:0.3

And, at least for me, going over a total strength of 1 would result in heavy artefacts (sweetspot round .75-.85)