The Lora is trained on MidJourney Images that use my "Personalised Style" + Vaporwave.

The Updated version SDXLv2 is cleaner and much easier to use.

Works without the trigger words ArsMJStyle, Vaporwave

Try lowering CFG if everything is too "melty"

I recommend ~1 for best "Vaporwave" results.

V1 Pony, but have effect on SDXL as well. Small dataset with stronger effect.

V2 have a little bit weaker effect , but is much more versatile .