A LoRA to enhance composition and lighting in anime-style samples, giving them a cinematic look.
There are two versions of this LoRA trained on the same data but with different parameters. Use any of these or use them together with any weight that works fine for you (0.5 may be enough).
Trigger words aren't required, but you may add these to strengthen effect: light particles, film grain, cinematic lighting, sidelighting, lens flare, depth of field
Description
FAQ
Comments (3)
Cinematic heavy makes a black string on some resolutions. Is it really something anime, or you just use it on anime checkpoins?
Here is a snapshot of training samples folder https://i.ibb.co/HGTwrbt/2024-01-06-09-30-51-cinema.png
The training was done on original SDXL 1.0 base model
Nice Lora! Ty ^^