animemixx

Details

Model description

LoRA Description:

This LoRA was created from stunning, manga-inspired images and is capable of reproducing beautiful colors, shadows, landscapes, and manga-style characters. Feel free to share your best creations—I may use them to develop an improved version, with credit to you, of course.

Technical Details for Training:
Optimizer: AdamW8
Scheduler: Cosine with Restart (3)
Learning Rate: 0.0003
Rank/Dim/Alpha: 16

Special thanks to Yume and Ngone for creating the wonderful images of the castle and cathedral.

Images made by this model