Training model: Animagine XL V3.1, https://civarchive.com/models/260267/animagine-xl-v3
Recommended Image Resolution: 1024+
Recommended Weight: 0.7 - 1.1
Prompts: long hair, white hair, twintails, hair rings, hair ornament, white eyelashes
Outfit: earrings, chinese clothes, two-tone dress, black dress, white dress, cleavage cutout, white thighhighs
Recommended to use these prompts also for anime image generation:
anime screencap, high quality anime 2d color, 2d, Accurate describe shapes, beautiful face, expressive eyes, natural lighting, light particles, cinematic angle, (masterpiece), (best quality), (ultra-detailed), very aesthetic, illustration, perfect composition, intricate details, moody lighting, light rays, depth of field, cowboy shot, cinematic angle, establish,
Recommended Setup: Steps: 20, cfg: 7.0, Sampler: Euler a
Use Animagine XL V3.0 for best results.
I'm still struggling to find high quality reference images for her clothes 💩
Description
I baked it longer this time, with 30 image repetitions. This time the images should look more similar to that of the game.
FAQ
Comments (3)
Somehow, I get a lot of arts with pink/purple eyes, even if I put (white eyes) in prompt. Well, I tried to put those eye colors on neg. prompt, but then I ended up with green/turquoise eyes. Overall new version felt better, but those eyes forced me to use previous one.
I do love this style. Are planning on adding an Pony version as well?
Favorite Jinhsi LoRA. Thank you so much for your work. The only problem, probably just for me, is her game outfit that try to appear when I don't want to. Especially her detachable collar. Negative prompt doesn't help. Otherwise, very good LoRA. Thanks.