米山舞风格 [米山 舞] Style
Vivid Impactful Style
An Artstyle Lora With High Compatibility
Important: Now, LoCon version is out!
Recommeneded prompts: Please check sample pictures.
Trigger word: vvi\(artstyle\), vvi\(face\),vvi\(eyes\),vvi\(oa\)
Note: You can use vvi\(artstyle\) to trigger the artsyle lora. If you want to generate portriats, I would recommand you use vvi\(face\),vvi\(eyes\) thses two trigger words to get a better performance. You can use vvi\(oa\) to trigger the offcial_art like style.
Recommeneded Weight: 0.6-1 (You can adjust it by your perference.)
CFG scale: 5.5-7
Check Point/Base Model recommendation: Anything V4.5, Dan Mix
Special thanks:
Artoria Caster (アルトリア・キャスター) – Lycoris
Atelier Shallie – Shallotte Elminus LoRA
(Lora usage in sample)
I am planning to release a LoHA version next week. No significant difference between LoCon and LoHa version. I will release a new version if I can get a larger training set.
How to use LoCon/LyCORIS? Please read instruction of a1111-sd-webui-locon.
What’s the different between v4g-lora and v4h-locon? Since I did not change the training set, I anticipate that they will have the same impact on your generated images. After hours of testing, I couldn’t find any significant difference between these two models. The v4h-locon may be slightly superior to the v4g-lora.
So, why am I still releasing this Locon? Because this v4h-locon has a file size advantage. It uses 100 mb less hard drive space,but still works great flawlessly.
v4g-lora和v4h-locon有什么不同？由于我没有改变训练集，我预计它们对您生成的图像将具有相同的影响。经过几个小时的测试，我无法找到这两个模型之间的任何显着差异。v4h-locon可能略优于v4g-lora。
那么，为什么我仍然要发布这个Locon模型呢？因为v4h-locon具有文件大小优势。Locon 模型比Lora模型小100 mb，但仍能完美无缺地运行。
Comparison (对比): All samples were generated with Anything V4.5 or Dan Mix V2.3 （底模：Anything V4.5/Dan Mix V2.3）
设计人生
