2025.01.25 更新基于2512版本01.25.2025 Update LoRA for Qwen Image 2512<a target="_blank" rel="ugc" href="https://www.modelscope.cn/models/fireicewolf/DukeG_Qwen_Light_And_Shadow_2512_Public/summary?version=20260113190656">Modelscope</a>地址Also on <a target="_blank" rel="ugc" href="https://www.modelscope.cn/models/fireicewolf/DukeG_Qwen_Light_And_Shadow_2512_Public/summary?version=20260113190656">Modelscope</a><hr />本LoRA旨在为Qwen Image模型带来更自然真实的光影效果。This LoRA aims to bring more natural and realistic light and shadow effects to Qwen Image.你可以使用ComyUI官方的Qwen Image文生图示例工作流。推荐权重：0.6-0.4.Recommand LoRA Weight：0.6-0.4.You can use the Qwen Image text-to-image workflow example provided by ComfyUI.如想要更好的生成效果，你可以在工作流中加入放大流程。For better generation results, you may add upscaling in your workflow.本LoRA由ModelScope社区提供训练算力支持。This LoRA was trained on the ModelScope.它同时也发布在<a target="_blank" rel="ugc" href="https://www.modelscope.cn/models/fireicewolf/DukeG_Qwen_Light_And_Shadow_Public">ModelScope</a>上。It also publiced on <a target="_blank" rel="ugc" href="https://www.modelscope.cn/models/fireicewolf/DukeG_Qwen_Light_And_Shadow_Public">ModelScope</a>.

DukeG Qwen Light & Shadow（DukeG 千问光影）

基于Qwen Image 2512重训练Re-trained base Qwen Image 2512

Files with SHA256: 8c37ee6be082ce2fe6af43f7c0221021759c5f0c5789f737b7148c3e44fac2d7