本页面包含 MOSS 和 Anima 两个版本，请选择对应的使用说明。 This page includes both MOSS and Anima versions. Please follow the instructions for your model._______________________________________________________________________________________<h2 id="recommended-settings">推荐参数/Recommended Settings</h2>FOR ANIMA:<ul><li>触发词/Trigger word: @yunsangstyle</li><li>权重/Recommended weight:0.7~1 (V2 Recommended:0.6-0.8)</li><li>Sampler: er_sde + beta OR euler + normal</li><li>CFG:4</li><li>STEPS:30</li></ul><h3 id="for-moss">FOR MOSS：</h3>base较好，但sft也尚可，base are better,but sft also ok.复杂场景，请提高CFG和步数，SFT例外。For complex scenes, increase the CFG and sampling steps,except when using SFT.<ul><li>触发词/Trigger word: @yunsangstyle</li><li>权重/Recommended weight:1~1.2</li><li>Sampler: er_sde + beta OR euler + simple</li><li>CFG:3-5</li><li>STEPS:30-50</li></ul><h3></h3>____________________________________________________________________________________<h2 id="moss">MOSS</h2><h3 id="v1">V1</h3><ul><li>和ANIMA同数据集训练，试了二十炉，不同配方，这一版是相对较好的，但感觉依然欠拟合，我大概会再作尝试。</li><li>Trained on the same dataset as ANIMA. I tried 20 runs with different training recipes, and this version is one of the better ones. It still feels underfit, though, so I’ll probably keep experimenting.</li><li>请测试，如果有什么问题，请留言告知。Please test it and let me know in the comments if you encounter any issues.</li></ul><h3 id="basesftbasethe-base-model-reproduces-the-style-best-overall.-sft-performs-better-in-certain-cases-but-base-is-more-faithful-to-the-original-art-style-in-most-situations.">base复原能力最好，但sft在某种情况上更好，但基本上在多数情况base更还原画风。The Base model reproduces the style best overall. SFT performs better in certain cases, but Base is more faithful to the original art style in most situations.</h3><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/3cd34b3e-58b7-4fc2-a89c-f5e3a27fdf90/original=true/3cd34b3e-58b7-4fc2-a89c-f5e3a27fdf90.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/52ecca41-24cf-4d0c-8d9a-f4b8c845de57/original=true/52ecca41-24cf-4d0c-8d9a-f4b8c845de57.jpeg" />____________________________________________________________________________________<h2 id="anima">ANIMA</h2><h3 id="v2:">v2:</h3><ul><li>target_res 1536，P-LORA ，修改了参数，指令遵从度上升，更好的泛化，光影提升、画面表现更好，增加少量数据集和局部参考。</li><li>Retrained at target_res 1536,P-LORA with adjusted training parameters, resulting in improved prompt adherence, better generalization, enhanced lighting and shadows, and higher overall image quality. A small amount of new training data and additional local reference images were also incorporated.</li></ul>_________________________________________________________________________________<h3 id="v1:">v1:</h3><ul><li>target_res 1536 重训，更好更细腻的纹理，和更好的泛化，皮肤质感提升极大。</li><li>Retrained with target_res 1536. Better and more refined textures, better generalization, and a huge improvement in skin texture.</li></ul><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/eff5b9e9-da8a-462a-9ed8-45760b8bed99/original=true/eff5b9e9-da8a-462a-9ed8-45760b8bed99.jpeg" /><h2 id="type-a-vs.-type-b">TYPE A vs. TYPE B</h2>typeA：更好的泛化，没有那么强烈的脸部和体型风格，配搭角色能够呈现出更多的脸型、体型、动作、构图，但相应皮肤质感没有B那么好。typeB：使用新画风3权重，老画风1权重的方式训练，强烈的风格，新画风更多的表现，如果泛化不好，请调低权重。如果你想要更强烈的风格，选B，如果想搭配更多的LORA使用，和更多样化，A会更好。<h2 id="type-a-vs.-type-b">TYPE A vs. TYPE B</h2>TYPE A — More flexible Better generalization with a less pronounced facial and body style. When paired with character LoRAs, it allows for a wider variety of face shapes, body types, poses, and compositions. The trade-off is that its skin texture is less refined than TYPE B’s.TYPE B — Stronger style Trained with a 3:1 weighting of the new art style to the old one, making the new style more prominent. If it doesn’t generalize well with your prompts, try lowering the LoRA weight.Which should you choose? Choose TYPE B for a stronger visual style. Choose TYPE A if you want to combine it with more LoRAs or create more varied results.

YUNSANG Style [MOSS // Anima]

ygstyle_v2.safetensors

Files with SHA256: 68753962e772797a5f90a365c7f52da781394126c9db79591db238c33c6096f3