Tensor.Art
在线生图

线条Line art

LORA
原创

353

必看:

1.本模型可以将白底黑线效果转为黑底金色线效果搭配ControlNet插件Canny (硬边缘)使用;

2.模型基于dreamshaper_8大模型训练,首推底模dreamshaper_8,也可搭配其余底模;

3.采样方法建议选用LCM,迭代步数5左右,提示词引导系数2-5(搭配lcm-lora-sdv1-5 使用时CFG可低一点,不使用LCM的LORA时可高一点)

4.建议搭配LCM的LORA同时使用,lora名称为:lcm-lora-sdv1-5

5.本模型权重本人常用1,也可根据需求自行设置

6.vae: 840000-ema-pruned | clip:2;

7.触发词:jxiantiao

Must see:

1. This model can convert the white background black line effect to a black background gold line effect, and use it with the ControlNet plugin Canny (hard edge);

2. The model is based on dreamshaper_ Training of 8 major models, starting with bottom model dreamshaper_ 8. It can also be paired with other bottom molds;

3. It is recommended to use LCM as the sampling method, with an iteration step of about 5 and a hint word guidance coefficient of 2-5 (CFG can be slightly lower when used with lcm lora sdv1-5, and slightly higher when using LORA without LCM)

4. It is recommended to use together with LCM's LORA, which is named lcm LORA sdv1-5

5. The weight of this model is commonly used by myself as 1, and can also be set according to needs

6. vae: 840000 ema run | clip: 2;

7. Trigger word: jxiantiao

Must see:

1. This model can convert the white background black line effect to a black background gold line effect, and use it with the ControlNet plugin Canny (hard edge);

2. The model is based on dreamshaper_ Training of 8 major models, starting with bottom model dreamshaper_ 8. It can also be paired with other bottom molds;

3. It is recommended to use LCM as the sampling method, with an iteration step of about 5 and a hint word guidance coefficient of 2-5 (CFG can be slightly lower when used with lcm lora sdv1-5, and slightly higher when using LORA without LCM)

4. It is recommended to use together with LCM's LORA, which is named lcm LORA sdv1-5

5. The weight of this model is commonly used by myself as 1, and can also be set according to needs

6. vae: 840000 ema run | clip: 2;

7. Trigger word: jxiantiao


版本详情

SD 1.5
必看: 1.本模型可以将白底黑线效果转为黑底金色线效果,搭配ControlNet插件Canny (硬边缘)使用,参数详见最下面配图; 2.模型基于dreamshaper_8大模型训练,首推底模dreamshaper_8,也可搭配其余底模; 3.采样方法建议选用LCM,迭代步数5左右,提示词引导系数2-5(搭配lcm-lora-sdv1-5 使用时CFG可低一点,不使用LCM的LORA时可高一点) 4.建议搭配LCM的LORA同时使用,lora名称为:lcm-lora-sdv1-5 5.本模型权重本人常用1,也可根据需求自行设置6.vae: 840000-ema-pruned | clip:2; 7.触发词:jxiantiao

项目权限

    使用权限

  • 使用时无需注明出处

  • 用于模型融合

  • 分享融合模型时使用不同的许可

    商用许可

  • 生成的图片用于商业用途

  • 作为图片生成服务来商用

  • 转售模型或出售融合模型

评论

相关帖子

描述想生成的画面,回车发送