A 2d anime LoRA model for vflower. base model: NovelAI 7G example images are generated with Anything-4.5 and MeinaPastelV2prompt: vflower, &lt;lora:vflower_v2:0.6&gt;, purple hairPS: the prompt "purp hair" is needed because some models seems to paint the hair pure white.le虚拟歌手 花花的模型 基于NovelAI 7G训练效果图使用Anything-4.5和MeinaPastelV2生成prompt: vflower, &lt;lora:vflower_v2:0.6&gt;, purple hair 我测试当中发现头发默认是白色的，所以需要加上"purple hair"这个prompt增强效果 V2 and V1 differences:<ul><li>less watermarks ( at least doesn't occur in conspicuous place )</li><li>needs less prompt</li><li>better effect?</li></ul>V2和V1的区别:<ul><li>去除了水印 (至少不会出现在显眼的地方)</li><li>需要更少的prompt</li><li>可能有更好的质量</li></ul>

vflower

Better? Only need vflower, &lt;lora:vflower:0.6&gt;, purple hair three prompts now, and remove the watermark i guess.

Files with SHA256: 4e0768ce18884c96ef2f47c797aee71882c0068935ddb67da6b5a5f96c43a271