################  AI大师工作流  #################说明：1、建议选择LCM或者lighting模型，这样生成图片速度会更快点，steps不用设那么高。2、controlnet模型可以选择sketch模型，这样既能能基本还原大致构图，又不至于被限制太死（毕竟你的草图画得如何，心理没点bi数-_-|| ？）。3、如果是想体现草稿图像为模型上的涂鸦或装饰，可以试试qrcode-monster模型。4、工作流默认会自动保存.obj模型文件在output文件夹中。5、如果想还原图片上的纹理颜色，可以使用该脚本:（<a target="_blank" rel="ugc" href="https://github.com/SunzeY/awesome-blender-script/blob/main/convert_to_std_obj.py），命令行中运行后进行素材贴图还原，然后导入到blender等专业建模软件中。">https://github.com/SunzeY/awesome-blender-script/blob/main/convert_to_std_obj.py），命令行中运行后进行素材贴图还原，然后导入到blender等专业建模软件中。</a>6、主要使用的TripoSR模型和相应节点( <a target="_blank" rel="ugc" href="https://github.com/flowtyone/ComfyUI-Flowty-TripoSR">https://github.com/flowtyone/ComfyUI-Flowty-TripoSR</a> )。7、抖音/B站/小红书/视频号等平台：AI大师工作流，欢迎关注。################  AI大师工作流  ##################Instructions:1, it is recommended to choose LCM or lighting model, so that the image generation speed will be faster, steps do not need to set so high.2, controlnet model can choose sketch model, which can basically restore the rough composition（After all, your sketches may not even be recognized by your mother. -_-||）.3, if you want to reflect the draft image as a doodle or decoration on the model, you can try qrcode-monster model.4, the workflow will automatically save the.obj model file in the output folder by default.5, If you want to restore the texture color on the image, you can use the script:(<a target="_blank" rel="ugc" href="https://github.com/SunzeY/awesome-blender-script/blob/main/convert_to_std_obj.py">https://github.com/SunzeY/awesome-blender-script/blob/main/convert_to_std_obj.py</a>), run in the command line to restore the material map, and then import it into blender or other professional 3D modeling software.6, the main use of TripoSR model and corresponding nodes(<a target="_blank" rel="ugc" href="https://github.com/flowtyone/ComfyUI-Flowty-TripoSR">https://github.com/flowtyone/ComfyUI-Flowty-TripoSR</a>).################  AI_master_workflow  ##################

Files with SHA256: 1d3aea26636eb4ecc702d09fd21b83983ae7a8c8df9fa1e48678e55dec7db8b0