<h2 id="introduction">Introduction</h2>This LoRA is trained on the illustration style of the artist <a target="_blank" rel="ugc" href="https://www.pixiv.net/users/10356511">ioriran04 / iorimiyakiyo</a>.<h3 id="usage">Usage</h3>Recommended LoRA strength: <code>1.0</code> Trigger word: <code>@shikuro (iorimiyakiyo)</code>In ComfyUI, escape the parentheses:<pre><code>@shikuro \(iorimiyakiyo\)</code></pre><h3 id="training-details">Training Details</h3><ul><li>Base model: Anima Base 1.0</li><li>Trainer: <code>sorryhyun/anima_lora v1.17.1</code></li><li>Dataset: 62 captioned images</li><li>LoRA rank / alpha: <code>32</code> / <code>128</code></li><li>Learning rate: <code>2e-5</code></li><li>Optimizer: AdamW</li><li>Scheduler: Cosine, with 5% warmup</li><li>Batch size: <code>1</code></li><li>Training duration: 40 epochs / 2480 steps</li><li>Precision: BF16</li><li>Resolution: Aspect-ratio buckets at approximately 1 megapixel</li><li>GPU: NVIDIA RTX 3090</li></ul><h3>简介</h3>这是基于 Anima Base 1.0 训练的 <a target="_blank" rel="ugc" href="https://www.pixiv.net/users/10356511">ioriran04 / iorimiyakiyo</a> 画风 LoRA。<h3>使用</h3>推荐 LoRA 权重： <code>1.0</code>激活词： <code>@shikuro (iorimiyakiyo)</code>在 ComfyUI 中需要转义括号：<pre><code>@shikuro \(iorimiyakiyo\)</code></pre><h3>训练参数</h3><ul><li>基础模型： Anima Base 1.0</li><li>训练工具： <code>sorryhyun/anima_lora v1.17.1</code></li><li>数据集： 62 张带 caption 的图片</li><li>LoRA rank / alpha： <code>32</code> / <code>128</code></li><li>学习率： <code>2e-5</code></li><li>优化器： AdamW</li><li>学习率调度： Cosine，5% warmup</li><li>Batch size： <code>1</code></li><li>训练量： 40 epochs / 2480 steps</li><li>训练精度： BF16</li><li>训练分辨率： 约 1 M 像素的宽高比分桶</li><li>显卡： NVIDIA RTX 3090</li></ul>

Files with SHA256: ae7090a6389360705121798a7cb0d3c2ca9a1c5941c61502e5dea765f9c47445