Added more samples to the dataset and gained more steps to train







New versions are coming soon!
V3 - SDXL Makima in photoshoot style, trained on 25000 steps LoRA AdamW fp16.
V2 - SDXL Makima in photoshoot style, trained on 10000 steps LoRA AdamW fp16.
V1 - SDXL Pony Makima in photoshoot style, trained on 4000+ steps LoRA AdamW fp16.
To prevent showing black screen, it's recommended to use SDXL Vae Fix FP16:
madebyollin/sdxl-vae-fp16-fix · Hugging Face
(big Thanks <3)
I don't know how to fix triple legs/bad hands/bad foot, maybe you can help me?
Also I don't know how to fix artifacts on faces, so should use CodeFormer or other face restoration methods.
Added more samples to the dataset and gained more steps to train
1. 转载模型仅供学习与交流分享,其版权及最终解释权归原作者。
2. 模型原作者如需认领模型,请通过官方渠道联系海艺AI工作人员进行认证。我们致力于保护每一位创作者的权益。 点击去认领
