Dataset:1024*768
BOTH FOR 2.5d Checkpoint and Realistic Checkpoint.
Weight : 0.8-1 for background ||||| 1-2 for characters.
This Lora is svg merge by two lora (Trained by partly common datasets) in order to great performance BOTH in Background and characters. No dataset pollution.
This lora performs well on landscapes and characters, but the training method is cumbersome, and it takes a lot of time to adjust the parameters without affecting the painting style.
