Trained on Hassaku, PixAI site, 11 images. It would seem to work great.

It's compatible with models like Koji v2.0 - v2.1, Dark Sushi, Galena REDUX, GrahamAingPikas Clear VAE2.3, etc.

I used ze default 0.7 LoRa strength, which works well enough. You may experiment if you feel like it. I used CFG Scale "7-8", again, feel free to experiment.

I'm using strong LoRa, "Add More Details - Detail Enhancer" ( https://pixai.art/model/1622983818871110869) for ze background as a default now, I'm not sure how good is this LoRa at generating rich sceneries or props by itself. Thumbnail model: Koji v2.1