<h2 id="asakura's-channel-style-lora-(il10-ver.)">🎨 Asakura's Channel Style LoRA (IL10 Ver.)</h2>[CN] 这是一个基于 Illustrious-XL v1.0 训练的风格向 LoRA，旨在还原 Channel (caststation) 独特的艺术笔触与色彩风格。 [EN] This is a style LoRA trained on Illustrious-XL v1.0, designed to capture the unique artistic brushwork and color palette of Channel (caststation).<hr /><h3 id="or-technical-insight">📊 技术可视化 | Technical Insight</h3>[CN] 为了证明模型的训练质量，我附带了 SVD (奇异值分解) 的可视化结果。 [EN] To demonstrate the training quality, I have included the SVD (Singular Value Decomposition) visualization.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/313ed25d-48a4-4d6b-9ece-7ad1643533d8/original=true/313ed25d-48a4-4d6b-9ece-7ad1643533d8.jpeg" />Technical Note: [CN] 观察 SVD 曲线可以发现，能量高度集中在前面的少数分量中。 [EN] The SVD curve shows that energy is highly concentrated in the first few components.<hr /><h3 id="or-features">🌟 模型特点 | Features</h3><ul><li>即插即用 | Plug and Play<ul><li>[CN] 无需触发词：载入 LoRA 即可生效。</li><li>[EN] No Trigger Word required: The style applies immediately upon loading.</li></ul></li><li>权重稳定 | Stable Weights<ul><li>[CN] 建议权重为 1.0，无需为了平衡画面频繁调整。</li><li>[EN] Recommended weight is 1.0. No constant tweaking is needed to balance the output.</li></ul></li></ul><hr /><h3 id="or-usage-tips">🚀 使用建议 | Usage Tips</h3><ul><li>底模 | Base Model: Illustrious-XL v1.0 &amp; variants.</li><li>权重 | Weight: <code>1.0</code> (Recommended) / <code>0.8 - 1.2</code>.</li><li>分辨率 | Resolution: 1024x1024+ (Hires. fix recommended).</li></ul><hr /><h3 id="or-disclaimer">📝 声明 | Disclaimer</h3>[CN] 本模型仅用于学术交流与艺术创作分享，请勿用于任何非法用途。如果你喜欢这个 LoRA，欢迎在 Civitai 下方分享你的作品！ [EN] This model is for academic exchange and artistic creation only. Please do not use it for any illegal purposes. If you enjoy this LoRA, feel free to share your generations below!

Files with SHA256: 50daa03debda606777e67ce5925317582ebc5d7a943bacef3b31efbc5fbe80c1