<h2 id="introduction">Introduction</h2>If you can only see the cover image, you may need to enable at least R-rated content in your Civitai content settings. Or check out the comparison image <a rel="ugc" href="https://ibb.co/mVwRqdHw">here</a>This is a Visual Concept Slider LoRA for Anima that controls the apparent opacity and thickness of hosiery. It works with pantyhose, tights, stockings, thigh-highs, and similar garments. Negative weights make the hosiery progressively thinner and more transparent, while positive weights make it denser and more opaque.<ul><li>✅ Works well with both SFW and NSFW subjects</li><li>✅ Supports different hosiery colors, such as red and yellow</li><li>✅ Plug and play; tested successfully on Anima Base and fine-tuned models such as WAI-Anima</li><li>⚠️ Preview version: Very high weights may slightly alter the composition (although no significant changes occurred in my testing) or shift the visual emphasis</li></ul><h2 id="usage">Usage</h2>Recommended weight: -1.0 to 1.0 (positive values make the hosiery thicker and more opaque; negative values make it thinner and more transparent)This LoRA does not require a trigger word.<h2 id="training-details">Training Details</h2><ul><li>Base model: Anima Base 1.0</li><li>Training tool: <code>sorryhyun/anima_lora v1.17.1</code> (with a custom Paired Visual Concept Slider loss function!)</li><li>Training dataset: 11 manually paired image sets</li><li>LoRA rank / alpha: 4 / 1</li><li>Learning rate: <code>2e-4</code></li><li>Optimizer: AdamW</li><li>Learning rate schedule: Constant learning rate with no warmup</li><li>Training duration: Approximately 22 epochs / 240 paired updates</li><li>Locality preservation weight: 0.5</li><li>Released checkpoint: Step 240</li><li>Training precision: BF16</li><li>Training resolution: Aspect-ratio buckets, with a target resolution of approximately <code>1024 × 896</code></li><li>GPU: NVIDIA GeForce RTX 3090</li></ul><h2>简介</h2>如果只能看到标题图，你可能需要在 Civitai 内容设置中打开至少 R 等级。或者在<a rel="ugc" href="https://ibb.co/mVwRqdHw">这里</a>查看对比图这是一个适用于 Anima 的 Visual Concept Slider LoRA，用于控制袜类的视觉透明度和厚度表现。它适用于裤袜、连裤袜、丝袜和长筒袜等，可以让袜子在负权重下逐渐变得轻薄、透肤，在正权重下逐渐变得更加致密和不透明。<ul><li>✅在 SFW 和 NSFW 主题下均工作良好</li><li>✅支持不同颜色的袜类（例如红色、黄色等）</li><li>✅即插即用，在 Anima Base 和 WAI-Anima 等微调上均测试有效</li><li>⚠️预览版：过高的权重可能会略微改变构图（在我的测试中，没有明显影响）或影响视觉重心</li></ul><h2>使用</h2>推荐权重：-1.0 - 1.0（正权重为变厚/不透明，负权重为变薄/透明）本 LoRA 没有任何激活词。<h2>训练参数</h2><ul><li>基础模型：Anima Base 1.0</li><li>训练工具：<code>sorryhyun/anima_lora v1.17.1</code>（自己编写了 Paired Visual Concept Slider Loss 函数！）</li><li>训练数据：11 组人工配对图片</li><li>LoRA rank / alpha：4 / 1</li><li>学习率：<code>2e-4</code></li><li>优化器：AdamW</li><li>学习率调度：恒定学习率，无预热</li><li>训练量：约 22 epochs / 240 paired updates</li><li>局部保持权重：0.5</li><li>发布版本：240-step</li><li>训练精度：BF16</li><li>训练分辨率：按长宽比分桶，目标分辨率约 <code>1024 × 896</code></li><li>训练显卡：NVIDIA GeForce RTX 3090</li></ul>

Files with SHA256: f45b4341dee82d5873b74cd319bf05e0f18c989a11948a04147ae31be80e4614