<a target="_blank" rel="ugc" href="https://ko-fi.com/zyd232">Buy me a coffee ☕</a><h3 id="features-ypj8q9v91">Features</h3><ol><li>Generate Detailed Hires Images</li><li>Edit Expressions - Freely adjust the expression through a sample image or parameters.</li><li>Face Detailer - Generate clear face when the face is too small.</li><li>Hand Fix - Fix the hands if necessary.</li></ol>　<h3 id="requirement-dyqkmxcvt">Requirement</h3><ul><li>Models<ol><li>Any Flux.1 unet model as the base model.\ComfyUI\models\unet\</li><li>VAE: <a target="_blank" rel="ugc" href="https://huggingface.co/black-forest-labs/FLUX.1-schnell/blob/main/ae.safetensors">ae.safetensors</a>.\ComfyUI\models\vae\</li><li>CLIP: <a target="_blank" rel="ugc" href="https://huggingface.co/lllyasviel/flux_text_encoders/blob/main/t5xxl_fp8_e4m3fn.safetensors">t5xxl_fp8_e4m3fn.safetensors</a>, <a target="_blank" rel="ugc" href="https://huggingface.co/lllyasviel/flux_text_encoders/blob/main/clip_l.safetensors">clip_l.safetensors</a>.\ComfyUI\models\clip\</li><li>Upscale:2K image (Default): <a target="_blank" rel="ugc" href="https://github.com/Phhofm/models/releases/tag/2xNomosUni_span_multijpg_ldl">2xNomosUni_span_multijpg_ldl</a>4K image (Optional): <a target="_blank" rel="ugc" href="https://github.com/Phhofm/models/releases/tag/4xNomos8kSCHAT">4xNomos8kSCHATL</a>.\ComfyUI\models\upscale_models\</li><li>Ultralytics: <a target="_blank" rel="ugc" href="https://huggingface.co/Bingsu/adetailer/blob/main/face_yolov8m.pt">face_</a><a target="_blank" rel="ugc" href="http://yolov8m.pt">yolov8m.pt</a>.\ComfyUI\models\ultralytics\bbox\</li><li>Sam: <a target="_blank" rel="ugc" href="https://huggingface.co/datasets/Gourieff/ReActor/blob/main/models/sams/sam_vit_b_01ec64.pth">sam_vit_b_01ec64.pth</a>.\ComfyUI\models\sams\</li><li>ControlNet:v1.0: <a target="_blank" rel="ugc" href="https://huggingface.co/Shakker-Labs/FLUX.1-dev-ControlNet-Union-Pro/blob/main/diffusion_pytorch_model.safetensors">FLUX.1-dev-ControlNet-Union-Pro</a>v1.1: <a target="_blank" rel="ugc" href="https://huggingface.co/jasperai/Flux.1-dev-Controlnet-Upscaler/blob/main/diffusion_pytorch_model.safetensors">Flux.1-dev-Controlnet-Upscaler</a>download and rename the file, put it into the path: .\ComfyUI\models\controlnet\Flux.1\</li></ol></li><li>Custom Nodes<ul><li>After importing the workflow for the first time, please use the "Install Missing Custom Nodes" button in <a target="_blank" rel="ugc" href="https://github.com/ltdrdata/ComfyUI-Manager">ComfyUI Manager</a> menu to get all the nodes needed.</li></ul></li></ul>　<h3 id="tips-n87ojz5h3">Tips</h3><ol><li>Change the latent width and height to your favorite, in Node "Empty Latent Size Picker", Group "Origin_Flux".</li><li>This workflow consumes huge VRAM, due to the ControlNet model for Flux. If you don't have enough VRAM, bypass the group called "Refine_Flux" to skip the ControlNet part. However, that means the images you get will not be as clear as the example images.</li><li>If you want higher resolution (4K) images, you can change the upscale model to <a target="_blank" rel="ugc" href="https://github.com/Phhofm/models/releases/tag/4xNomos8kSCHAT">4xNomos8kSCHAT-L</a>.</li></ol>　__________________________________________________________________<a target="_blank" rel="ugc" href="https://afdian.com/a/zyd232">支持我的工作 ☕</a><h3 id="-jlac75f30">特点</h3><ol><li>生成高清且细节丰富的图片</li><li>编辑表情 - 通过样本图像或参数，自由调整表情。</li><li>面部修复 - 在面部太小时生成清晰的面部。</li><li>手部修复 - 在需要时修复手部。</li></ol>　<h3 id="-x7dyvu9vh">需求</h3><ul><li>模型<ol><li>任何 Flux.1 unet 作为底模.\ComfyUI\models\unet\</li><li>VAE: <a target="_blank" rel="ugc" href="https://huggingface.co/black-forest-labs/FLUX.1-schnell/blob/main/ae.safetensors">ae.safetensors</a>.\ComfyUI\models\vae\</li><li>CLIP: <a target="_blank" rel="ugc" href="https://huggingface.co/lllyasviel/flux_text_encoders/blob/main/t5xxl_fp8_e4m3fn.safetensors">t5xxl_fp8_e4m3fn.safetensors</a>, <a target="_blank" rel="ugc" href="https://huggingface.co/lllyasviel/flux_text_encoders/blob/main/clip_l.safetensors">clip_l.safetensors</a>.\ComfyUI\models\clip\</li><li>Upscale:2K 图片（默认）: <a target="_blank" rel="ugc" href="https://github.com/Phhofm/models/releases/tag/2xNomosUni_span_multijpg_ldl">2xNomosUni_span_multijpg_ldl</a>4K 图片（可选）: <a target="_blank" rel="ugc" href="https://github.com/Phhofm/models/releases/tag/4xNomos8kSCHAT">4xNomos8kSCHATL</a>.\ComfyUI\models\upscale_models\</li><li>Ultralytics: <a target="_blank" rel="ugc" href="https://huggingface.co/Bingsu/adetailer/blob/main/face_yolov8m.pt">face_</a><a target="_blank" rel="ugc" href="http://yolov8m.pt">yolov8m.pt</a>.\ComfyUI\models\ultralytics\bbox\</li><li>Sam: <a target="_blank" rel="ugc" href="https://huggingface.co/datasets/Gourieff/ReActor/blob/main/models/sams/sam_vit_b_01ec64.pth">sam_vit_b_01ec64.pth</a>.\ComfyUI\models\sams\</li><li>ControlNet:v1.0: <a target="_blank" rel="ugc" href="https://huggingface.co/Shakker-Labs/FLUX.1-dev-ControlNet-Union-Pro/blob/main/diffusion_pytorch_model.safetensors">FLUX.1-dev-ControlNet-Union-Pro</a>v1.1: <a target="_blank" rel="ugc" href="https://huggingface.co/jasperai/Flux.1-dev-Controlnet-Upscaler/blob/main/diffusion_pytorch_model.safetensors">Flux.1-dev-Controlnet-Upscaler</a>下载并重命名文件, 将其放在:.\ComfyUI\models\controlnet\Flux.1\</li></ol></li><li>自定义节点<ul><li>首次导入工作流程后，请使用 <a target="_blank" rel="ugc" href="https://github.com/ltdrdata/ComfyUI-Manager">ComfyUI Manager</a> 菜单中的 “Install Missing Custom Nodes” 按钮获取所需的所有节点。</li></ul></li></ul>　<h3 id="-qf4uvfql2">提示</h3><ol><li>在 “Origin_Flux ”组中的 “Empty Latent Size Picker”节点里，将宽度和高度改为你想要值。</li><li>由于 Flux 的 ControlNet 模型，此工作流会消耗大量显存。如果你没有足够的显存，可以把 “Refine_Flux ”组设为 bypass，以跳过 ControlNet 部分。不过，这意味着你获得的图像不会像示例图片那样清晰。</li><li>如果想要更高分辨率（4K）的图片，可以把Upscale模型换成 <a target="_blank" rel="ugc" href="https://github.com/Phhofm/models/releases/tag/4xNomos8kSCHAT">4xNomos8kSCHAT-L</a>。</li></ol>

Files with SHA256: daf213eb62dac34cf050e57c35c58cdf22ed177cc8a6011cb212ad0599649457