chokopaii style<a target="_blank" rel="ugc" href="https://x.com/ch0k0paii">https://x.com/ch0k0paii</a><h1 id="v2">V2</h1>神童版本已上传，不知道为什么拟合的不是很好，以后再做尝试吧<h1 id="v1">V1</h1>初次尝试了一下最近的开源的MOSS模型，感觉有点欠拟合，而且生图速度确实太感人了，暂时还是留在阿尼玛吧！不过好处是它真的认识很多角色，在角色细节的还原上很强，甚至会写文字以后moss会和其他模型的lora单独分开English ver:I gave the recently released open-source MOSS model a try. It feels a bit underfitted, and the image generation speed is—well, let’s just say it’s "impressively" slow. For now, I think I’ll stick with Anima!On the plus side, it’s familiar with a huge number of characters and does an excellent job of capturing specific character details—it can even render text.Moving forward, I’ll be keeping MOSS-related LoRAs separate from those for other models.这是Adamw8bit炼的，晚点会上传一个神童炼的版本

Files with SHA256: a97aa4bc7b4ca95f4ec2b5ef7093c5739b8636ead52b66bfb89d062e15c21e75