<h3>模型简介</h3>本 LoRA 旨在完美还原 Pixiv 大佬 曦晖貔翊（<a target="_blank" rel="ugc" href="https://www.pixiv.net/users/110653205">曦晖貔翊 - pixiv</a>）的精美画风。<h3>生成参数与示例图说明</h3>如果您参考了示例图的生成数据，会发现采样步数（Steps）和 CFG 值都设置得非常低。这是因为示例图在生成时，搭配使用了特定的加速与细节增强 LoRA：<ul><li>加速 LoRA: <code>dmd2_sdxl_4step_lora</code> (<a target="_blank" rel="ugc" href="https://civarchive.com/models/1608870/dmd2-speed-lora-sdxl-pony-illustrious?modelVersionId=1820705">DMD2 Speed LoRA [SDXL, Pony, Illustrious] - dmd2_sdxl_4step_lora_fp32 | Stable Diffusion XL LoRA | Civitai</a>)</li><li>细节 LoRA: <code>NOOB_vp1_detailer_by_volnovik_v1</code> (<a target="_blank" rel="ugc" href="https://civarchive.com/models/987736/noobai-xl-detailer?modelVersionId=1209350">NoobAI-XL Detailer - V-Pred1_V1 | NoobAI LoRA | Civitai</a>)</li></ul>注：如果您在不搭配 DMD2 加速 LoRA 的情况下进行生成，请务必将步数和 CFG 值调回常规的 NoobAI-XL 参数标准。<h3 id="model-overview">Model Overview</h3>This LoRA replicates the beautiful and distinctive art style of the Pixiv artist 曦晖貔翊 (<a target="_blank" rel="ugc" href="https://www.pixiv.net/users/110653205">曦晖貔翊 - pixiv</a>).<h3 id="generation-notes-and-sample-image-settings">Generation Notes &amp; Sample Image Settings</h3>If you check the generation data of the sample images, you will notice unusually low sampling steps and a very low CFG scale. This is because the showcase images were generated in conjunction with specific acceleration and detailing LoRAs:<ul><li>Speed LoRA: <code>dmd2_sdxl_4step_lora</code> (<a target="_blank" rel="ugc" href="https://civarchive.com/models/1608870/dmd2-speed-lora-sdxl-pony-illustrious?modelVersionId=1820705">DMD2 Speed LoRA [SDXL, Pony, Illustrious] - dmd2_sdxl_4step_lora_fp32 | Stable Diffusion XL LoRA | Civitai</a>)</li><li>Detailer LoRA: <code>NOOB_vp1_detailer_by_volnovik_v1</code> (<a target="_blank" rel="ugc" href="https://civarchive.com/models/987736/noobai-xl-detailer?modelVersionId=1209350">NoobAI-XL Detailer - V-Pred1_V1 | NoobAI LoRA | Civitai</a>)</li></ul>Note: If you are generating without the DMD2 Speed LoRA, please remember to adjust your steps and CFG scale back to standard NoobAI-XL parameters.免责声明与使用条款<ol><li>公益性质：本模型为免费开源的公益项目，仅供个人学习、研究及非商业性创作使用。作者不收取任何费用，也未授权任何第三方进行售卖。</li><li>内容合规：请勿使用本模型生成违反法律法规、侵犯他人权益、色情暴力或具有歧视性的内容。因不当使用产生的一切法律责任由使用者自行承担，模型作者不承担任何连带责任。</li><li>知识产权：本模型基于Cat Tower (NoobAI XL checkpoint) 底膜训练，训练数据来源于公开网络。如侵犯了您的合法权益，请及时联系作者进行下架处理。</li><li>无担保声明：本模型按“现状”提供，作者不对模型的特定用途适用性、稳定性或生成结果做任何明示或暗示的保证。</li><li>禁止事项：严禁将本模型用于诈骗、深度伪造（Deepfake）诽谤他人、或作为付费服务/商品的核心组件进行牟利。</li><li>任何基于本模型的融合（Merge）、微调或衍生模型，也必须严格遵守非商业用途和免费开源的原则，不得更改版权协议。</li></ol>Disclaimer &amp; Terms of Use<ol><li>Non-Profit / Public Welfare: This model is released as a free, open-source public welfare project. It is intended solely for personal learning, research, and non-commercial creative use. The author charges no fees and has not authorized any third-party sales.</li><li>Content Compliance: Do NOT use this model to generate content that violates laws, infringes on others' rights, is pornographic/violent, or discriminatory. Users bear full legal responsibility for any misuse; the model author assumes no liability.</li><li>Intellectual Property: This model is trained based on the bottom layer of Cat Tower (NoobAI XL checkpoint), and the training data is sourced from the public network. If your rights have been infringed upon, please contact the author promptly for removal.</li><li>No Warranty: This model is provided "AS IS" without warranty of any kind, express or implied, regarding fitness for a particular purpose or output quality.</li><li>Prohibited Uses: Strictly prohibited for fraud, defamatory deepfakes, or as a core component of paid services/products for profit.</li><li>Any merged, fine-tuned, or derivative models based on this model must also strictly adhere to the non-commercial and free open-source principles, and the license terms cannot be altered.</li></ol>

artstyle:xihuipiyi

模型简介本 LoRA 旨在完美还原 Pixiv 大佬 曦晖貔翊（<a target="_blank" rel="ugc" href="https://www.pixiv.net/users/110653205">曦晖貔翊 - pixiv</a>）的精美画风。生成参数与示例图说明如果您参考了示例图的生成数据，会发现采样步数（Steps）和 CFG 值都设置得非常低。这是因为示例图在生成时，搭配使用了特定的加速与细节增强 LoRA：<ul><li>加速 LoRA: <code>dmd2_sdxl_4step_lora</code> (<a target="_blank" rel="ugc" href="https://civitai.red/models/1608870/dmd2-speed-lora-sdxl-pony-illustrious?modelVersionId=1820705">DMD2 Speed LoRA [SDXL, Pony, Illustrious] - dmd2_sdxl_4step_lora_fp32 | Stable Diffusion XL LoRA | Civitai</a>)</li><li>细节 LoRA: <code>NOOB_vp1_detailer_by_volnovik_v1</code> (<a target="_blank" rel="ugc" href="https://civitai.red/models/987736/noobai-xl-detailer?modelVersionId=1209350">NoobAI-XL Detailer - V-Pred1_V1 | NoobAI LoRA | Civitai</a>)</li></ul>注：如果您在不搭配 DMD2 加速 LoRA 的情况下进行生成，请务必将步数和 CFG 值调回常规的 NoobAI-XL 参数标准。Model OverviewThis LoRA replicates the beautiful and distinctive art style of the Pixiv artist 曦晖貔翊 (<a target="_blank" rel="ugc" href="https://www.pixiv.net/users/110653205">曦晖貔翊 - pixiv</a>).Generation Notes &amp; Sample Image SettingsIf you check the generation data of the sample images, you will notice unusually low sampling steps and a very low CFG scale. This is because the showcase images were generated in conjunction with specific acceleration and detailing LoRAs:<ul><li>Speed LoRA: <code>dmd2_sdxl_4step_lora</code> (<a target="_blank" rel="ugc" href="https://civitai.red/models/1608870/dmd2-speed-lora-sdxl-pony-illustrious?modelVersionId=1820705">DMD2 Speed LoRA [SDXL, Pony, Illustrious] - dmd2_sdxl_4step_lora_fp32 | Stable Diffusion XL LoRA | Civitai</a>)</li><li>Detailer LoRA: <code>NOOB_vp1_detailer_by_volnovik_v1</code> (<a target="_blank" rel="ugc" href="https://civitai.red/models/987736/noobai-xl-detailer?modelVersionId=1209350">NoobAI-XL Detailer - V-Pred1_V1 | NoobAI LoRA | Civitai</a>)</li></ul>Note: If you are generating without the DMD2 Speed LoRA, please remember to adjust your steps and CFG scale back to standard NoobAI-XL parameters.免责声明与使用条款<ol><li>公益性质：本模型为免费开源的公益项目，仅供个人学习、研究及非商业性创作使用。作者不收取任何费用，也未授权任何第三方进行售卖。</li><li>内容合规：请勿使用本模型生成违反法律法规、侵犯他人权益、色情暴力或具有歧视性的内容。因不当使用产生的一切法律责任由使用者自行承担，模型作者不承担任何连带责任。</li><li>知识产权：本模型基于Cat Tower (NoobAI XL checkpoint) 底膜训练，训练数据来源于公开网络。如侵犯了您的合法权益，请及时联系作者进行下架处理。</li><li>无担保声明：本模型按“现状”提供，作者不对模型的特定用途适用性、稳定性或生成结果做任何明示或暗示的保证。</li><li>禁止事项：严禁将本模型用于诈骗、深度伪造（Deepfake）诽谤他人、或作为付费服务/商品的核心组件进行牟利。</li><li>任何基于本模型的融合（Merge）、微调或衍生模型，也必须严格遵守非商业用途和免费开源的原则，不得更改版权协议。</li></ol>Disclaimer &amp; Terms of Use<ol><li>Non-Profit / Public Welfare: This model is released as a free, open-source public welfare project. It is intended solely for personal learning, research, and non-commercial creative use. The author charges no fees and has not authorized any third-party sales.</li><li>Content Compliance: Do NOT use this model to generate content that violates laws, infringes on others' rights, is pornographic/violent, or discriminatory. Users bear full legal responsibility for any misuse; the model author assumes no liability.</li><li>Intellectual Property: This model is trained based on the bottom layer of Cat Tower (NoobAI XL checkpoint), and the training data is sourced from the public network. If your rights have been infringed upon, please contact the author promptly for removal.</li><li>No Warranty: This model is provided "AS IS" without warranty of any kind, express or implied, regarding fitness for a particular purpose or output quality.</li><li>Prohibited Uses: Strictly prohibited for fraud, defamatory deepfakes, or as a core component of paid services/products for profit.</li><li>Any merged, fine-tuned, or derivative models based on this model must also strictly adhere to the non-commercial and free open-source principles, and the license terms cannot be altered.</li></ol>

Files with SHA256: 01f15c7fef3637eb77ead1736151c3f3d8e8206548b63ed9eefd55a5f6ae2383