{"id":31981,"date":"2025-07-01T16:59:03","date_gmt":"2025-07-01T08:59:03","guid":{"rendered":"https:\/\/www.kdjingpai.com\/?p=31981"},"modified":"2025-07-01T17:01:18","modified_gmt":"2025-07-01T09:01:18","slug":"nunchaku","status":"publish","type":"post","link":"https:\/\/www.kdjingpai.com\/en\/nunchaku\/","title":{"rendered":"Nunchaku\uff1a\u9ad8\u6548\u8fd0\u884cFLUX.1 \u548c SANA 4\u4f4d\u91cf\u5316\u6a21\u578b\u7684\u63a8\u7406\u5de5\u5177"},"content":{"rendered":"<p>Nunchaku \u662f\u7531 MIT HAN Lab \u5f00\u53d1\u7684\u4e00\u4e2a\u5f00\u6e90\u63a8\u7406\u5f15\u64ce\uff0c\u4e13\u6ce8\u4e8e\u9ad8\u6548\u8fd0\u884c4\u4f4d\u91cf\u5316\u7684\u6269\u6563\u6a21\u578b\u3002\u5b83\u901a\u8fc7 SVDQuant \u6280\u672f\uff0c\u5c06\u6a21\u578b\u7684\u6743\u91cd\u548c\u6fc0\u6d3b\u91cf\u5316\u4e3a4\u4f4d\uff0c\u663e\u8457\u964d\u4f4e\u5185\u5b58\u5360\u7528\u548c\u63a8\u7406\u5ef6\u8fdf\u3002Nunchaku \u652f\u6301 <a href=\"https:\/\/www.kdjingpai.com\/de\/manxuebanflux1chan\/\">FLUX<\/a> \u548c SANA \u7b49\u6269\u6563\u6a21\u578b\uff0c\u9002\u7528\u4e8e\u56fe\u50cf\u751f\u6210\u3001\u7f16\u8f91\u548c\u8f6c\u6362\u7b49\u4efb\u52a1\u3002\u5b83\u4e0e <a href=\"https:\/\/www.kdjingpai.com\/de\/comfyui\/\">ComfyUI<\/a> \u96c6\u6210\uff0c\u63d0\u4f9b\u7528\u6237\u53cb\u597d\u7684\u64cd\u4f5c\u754c\u9762\uff0c\u9002\u5408\u7814\u7a76\u4eba\u5458\u548c\u5f00\u53d1\u8005\u5728\u4f4e\u8d44\u6e90\u8bbe\u5907\u4e0a\u8fd0\u884c\u590d\u6742\u6a21\u578b\u3002\u9879\u76ee\u5728 GitHub \u4e0a\u5f00\u6e90\uff0c\u793e\u533a\u6d3b\u8dc3\uff0c\u63d0\u4f9b\u8be6\u7ec6\u6587\u6863\u548c\u793a\u4f8b\u811a\u672c\uff0c\u65b9\u4fbf\u7528\u6237\u5feb\u901f\u4e0a\u624b\u3002<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-31982\" title=\"Nunchaku\uff1a\u9ad8\u6548\u8fd0\u884c4\u4f4d\u91cf\u5316\u7684\u6269\u6563\u6a21\u578b\u63a8\u7406\u5de5\u5177-1\" src=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/07\/bb67e43698d99e5.png\" alt=\"Nunchaku\uff1a\u9ad8\u6548\u8fd0\u884c4\u4f4d\u91cf\u5316\u7684\u6269\u6563\u6a21\u578b\u63a8\u7406\u5de5\u5177-1\" width=\"1045\" height=\"661\" srcset=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/07\/bb67e43698d99e5.png 1045w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/07\/bb67e43698d99e5-18x12.png 18w\" sizes=\"auto, (max-width: 1045px) 100vw, 1045px\" \/><\/p>\n<p>&nbsp;<\/p>\n<h2>\u529f\u80fd\u5217\u8868<\/h2>\n<ul>\n<li>\u652f\u63014\u4f4d\u91cf\u5316\u6269\u6563\u6a21\u578b\u63a8\u7406\uff0c\u5185\u5b58\u5360\u7528\u964d\u4f4e3.6\u500d\uff0c\u901f\u5ea6\u63d0\u5347\u6700\u9ad88.7\u500d\u3002<\/li>\n<li>\u517c\u5bb9 FLUX.1 \u7cfb\u5217\u6a21\u578b\uff0c\u5305\u62ec FLUX.1-dev\u3001FLUX.1-schnell \u548c FLUX.1-tools\u3002<\/li>\n<li>\u63d0\u4f9b\u6587\u672c\u5230\u56fe\u50cf\u3001\u8349\u56fe\u5230\u56fe\u50cf\u3001\u6df1\u5ea6\u56fe\/\u8fb9\u7f18\u56fe\u5230\u56fe\u50cf\u3001\u56fe\u50cf\u4fee\u590d\u7b49\u591a\u79cd\u751f\u6210\u4efb\u52a1\u3002<\/li>\n<li>\u652f\u6301 LoRA \u52a0\u8f7d\uff0c\u5141\u8bb8\u7528\u6237\u52a0\u8f7d\u81ea\u5b9a\u4e49\u6a21\u578b\u4ee5\u589e\u5f3a\u751f\u6210\u6548\u679c\u3002<\/li>\n<li>\u96c6\u6210 ComfyUI\uff0c\u63d0\u4f9b\u8282\u70b9\u5f0f\u64cd\u4f5c\u754c\u9762\uff0c\u7b80\u5316\u5de5\u4f5c\u6d41\u914d\u7f6e\u3002<\/li>\n<li>\u652f\u6301\u81ea\u7136\u8bed\u8a00\u56fe\u50cf\u7f16\u8f91\uff0c\u901a\u8fc7 FLUX.1-Kontext-dev \u5b9e\u73b0\u6587\u672c\u9a71\u52a8\u7684\u56fe\u50cf\u4fee\u6539\u3002<\/li>\n<li>\u63d0\u4f9b FP16 \u7cbe\u5ea6\u652f\u6301\uff0c\u4f18\u5316\u6a21\u578b\u6027\u80fd\u548c\u751f\u6210\u8d28\u91cf\u3002<\/li>\n<li>\u63d0\u4f9b PuLID v0.9.1 \u96c6\u6210\uff0c\u652f\u6301\u81ea\u5b9a\u4e49\u6a21\u578b\u8def\u5f84\u548c timestep \u63a7\u5236\u3002<\/li>\n<\/ul>\n<h2>\u4f7f\u7528\u5e2e\u52a9<\/h2>\n<h3>\u5b89\u88c5\u6d41\u7a0b<\/h3>\n<p>Nunchaku \u7684\u5b89\u88c5\u9700\u8981\u786e\u4fdd\u7cfb\u7edf\u6ee1\u8db3\u57fa\u672c\u8981\u6c42\uff0c\u5305\u62ec PyTorch\u3001CUDA \u548c\u7279\u5b9a\u7f16\u8bd1\u5668\u3002\u4ee5\u4e0b\u662f\u8be6\u7ec6\u7684\u5b89\u88c5\u6b65\u9aa4\uff0c\u9002\u7528\u4e8e Linux \u548c Windows \u7cfb\u7edf\u3002<\/p>\n<h4>\u73af\u5883\u51c6\u5907<\/h4>\n<ol>\n<li><strong>\u5b89\u88c5 PyTorch<\/strong>\uff1a\u786e\u4fdd\u5b89\u88c5 PyTorch 2.6 \u6216\u66f4\u9ad8\u7248\u672c\u3002\u4f8b\u5982\uff0c\u4f7f\u7528\u4ee5\u4e0b\u547d\u4ee4\u5b89\u88c5 PyTorch 2.6\uff1a\n<pre><code>pip install torch==2.6 torchvision==0.21 torchaudio==2.6\r\n<\/code><\/pre>\n<p>\u5982\u679c\u4f7f\u7528 Blackwell GPU\uff08\u5982 RTX 50 \u7cfb\u5217\uff09\uff0c\u9700\u8981 PyTorch 2.7 \u53ca CUDA 12.8 \u6216\u66f4\u9ad8\u7248\u672c\u3002<\/li>\n<li><strong>\u68c0\u67e5 CUDA \u7248\u672c<\/strong>\uff1aLinux \u9700\u8981 CUDA 12.2 \u53ca\u4ee5\u4e0a\uff0cWindows \u9700\u8981 CUDA 12.6 \u53ca\u4ee5\u4e0a\u3002<\/li>\n<li><strong>\u5b89\u88c5\u7f16\u8bd1\u5668<\/strong>\uff1a\n<ul>\n<li>Linux\uff1a\u786e\u4fdd\u5b89\u88c5\u00a0<code>gcc\/g++<\/code>\u00a011 \u6216\u66f4\u9ad8\u7248\u672c\uff0c\u53ef\u901a\u8fc7 Conda \u5b89\u88c5\uff1a\n<pre><code>conda install -c conda-forge gcc gxx\r\n<\/code><\/pre>\n<\/li>\n<li>Windows\uff1a\u5b89\u88c5\u6700\u65b0\u7248\u672c\u7684 Visual Studio\uff0c\u786e\u4fdd\u5305\u542b C++ \u5f00\u53d1\u7ec4\u4ef6\u3002<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u5b89\u88c5\u4f9d\u8d56<\/strong>\uff1a\u5b89\u88c5\u5fc5\u8981\u7684 Python \u5305\uff1a\n<pre><code>pip install ninja wheel diffusers transformers accelerate sentencepiece protobuf huggingface_hub\r\n<\/code><\/pre>\n<p>\u5982\u679c\u9700\u8981\u8fd0\u884c Gradio \u6f14\u793a\uff0c\u989d\u5916\u5b89\u88c5\uff1a<\/p>\n<pre><code>pip install peft opencv-python gradio spaces\r\n<\/code><\/pre>\n<\/li>\n<\/ol>\n<h4>\u5b89\u88c5 Nunchaku<\/h4>\n<ol>\n<li><strong>\u4ece Hugging Face \u5b89\u88c5<\/strong>\uff1a\u6839\u636e Python \u548c PyTorch \u7248\u672c\u9009\u62e9\u5408\u9002\u7684 wheel \u6587\u4ef6\u3002\u4f8b\u5982\uff0cPython 3.11 \u548c PyTorch 2.6\uff1a\n<pre><code>pip install https:\/\/huggingface.co\/mit-han-lab\/nunchaku\/resolve\/main\/nunchaku-0.2.0+torch2.6-cp311-cp311-linux_x86_64.whl\r\n<\/code><\/pre>\n<p>Windows \u7528\u6237\u53ef\u4f7f\u7528\uff1a<\/p>\n<pre><code>pip install https:\/\/huggingface.co\/mit-han-lab\/nunchaku\/resolve\/main\/nunchaku-0.1.4+torch2.6-cp312-cp312-win_amd64.whl\r\n<\/code><\/pre>\n<\/li>\n<li><strong>\u4ece\u6e90\u7801\u5b89\u88c5<\/strong>\uff08\u53ef\u9009\uff09\uff1a\n<pre><code>git clone https:\/\/github.com\/mit-han-lab\/nunchaku.git\r\ncd nunchaku\r\ngit submodule init\r\ngit submodule update\r\nexport NUNCHAKU_INSTALL_MODE=ALL\r\npython setup.py develop\r\n<\/code><\/pre>\n<p>\u6ce8\u610f\uff1a\u8bbe\u7f6e\u00a0<code>NUNCHAKU_INSTALL_MODE=ALL<\/code>\u00a0\u786e\u4fdd\u751f\u6210\u7684 wheel \u6587\u4ef6\u517c\u5bb9\u4e0d\u540c GPU \u67b6\u6784\u3002<\/li>\n<\/ol>\n<h4>ComfyUI \u96c6\u6210<\/h4>\n<ol>\n<li><strong>\u5b89\u88c5 ComfyUI<\/strong>\uff1a\u5982\u679c\u5c1a\u672a\u5b89\u88c5 ComfyUI\uff0c\u53ef\u901a\u8fc7\u4ee5\u4e0b\u547d\u4ee4\u5b89\u88c5\uff1a\n<pre><code>git clone https:\/\/github.com\/comfyanonymous\/ComfyUI.git\r\ncd ComfyUI\r\npip install -r requirements.txt\r\n<\/code><\/pre>\n<\/li>\n<li><strong>\u5b89\u88c5 ComfyUI-nunchaku \u63d2\u4ef6<\/strong>\uff1a\n<pre><code>pip install comfy-cli\r\ncomfy install\r\ncomfy node registry-install ComfyUI-nunchaku\r\n<\/code><\/pre>\n<p>\u6216\u8005\u901a\u8fc7 ComfyUI \u7684 Custom Nodes Manager \u641c\u7d22\u5e76\u5b89\u88c5\u00a0<code>ComfyUI-nunchaku<\/code>\u3002<\/li>\n<li><strong>\u4e0b\u8f7d\u6a21\u578b<\/strong>\uff1a\u5c06\u5fc5\u8981\u7684\u6a21\u578b\u6587\u4ef6\u4e0b\u8f7d\u5230\u6307\u5b9a\u76ee\u5f55\u3002\u4f8b\u5982\uff1a\n<pre><code>huggingface-cli download comfyanonymous\/flux_text_encoders clip_l.safetensors --local-dir models\/text_encoders\r\nhuggingface-cli download comfyanonymous\/flux_text_encoders t5xxl_fp16.safetensors --local-dir models\/text_encoders\r\nhuggingface-cli download black-forest-labs\/FLUX.1-schnell ae.safetensors --local-dir models\/vae\r\n<\/code><\/pre>\n<\/li>\n<li><strong>\u590d\u5236\u5de5\u4f5c\u6d41<\/strong>\uff1a\u5c06\u793a\u4f8b\u5de5\u4f5c\u6d41\u590d\u5236\u5230 ComfyUI \u76ee\u5f55\uff1a\n<pre><code>mkdir -p user\/default\/example_workflows\r\ncp custom_nodes\/nunchaku_nodes\/example_workflows\/* user\/default\/example_workflows\/\r\n<\/code><\/pre>\n<\/li>\n<li><strong>\u8fd0\u884c ComfyUI<\/strong>\uff1a\n<pre><code>cd ComfyUI\r\npython main.py\r\n<\/code><\/pre>\n<\/li>\n<\/ol>\n<h3>\u4e3b\u8981\u529f\u80fd\u64cd\u4f5c<\/h3>\n<p>Nunchaku \u63d0\u4f9b\u591a\u79cd\u56fe\u50cf\u751f\u6210\u548c\u7f16\u8f91\u529f\u80fd\uff0c\u4ee5\u4e0b\u662f\u5177\u4f53\u64cd\u4f5c\u65b9\u6cd5\u3002<\/p>\n<h4>\u6587\u672c\u5230\u56fe\u50cf\u751f\u6210<\/h4>\n<ol>\n<li>\u6253\u5f00 ComfyUI\uff0c\u52a0\u8f7d Nunchaku \u5de5\u4f5c\u6d41\uff08\u5982\u00a0<code>nunchaku-flux.1-dev-pulid.json<\/code>\uff09\u3002<\/li>\n<li>\u5728\u5de5\u4f5c\u6d41\u4e2d\u9009\u62e9\u00a0<code>Nunchaku Text Encoder Loader<\/code>\uff0c\u8bbe\u7f6e\u00a0<code>use_4bit_t5=True<\/code>\u00a0\u4ee5\u4f7f\u7528 4 \u4f4d\u91cf\u5316\u7684 T5 \u6587\u672c\u7f16\u7801\u5668\uff0c\u8282\u7701\u5185\u5b58\u3002<\/li>\n<li>\u6307\u5b9a\u00a0<code>t5_min_length=512<\/code>\u00a0\u4ee5\u63d0\u9ad8\u56fe\u50cf\u8d28\u91cf\u3002<\/li>\n<li>\u8f93\u5165\u6587\u672c\u63d0\u793a\uff0c\u4f8b\u5982\u201ca futuristic city at sunset\u201d\uff0c\u8c03\u6574\u751f\u6210\u53c2\u6570\uff08\u5982\u91c7\u6837\u6b65\u6570\u3001\u5f15\u5bfc\u5c3a\u5ea6\uff09\u3002<\/li>\n<li>\u8fd0\u884c\u5de5\u4f5c\u6d41\uff0c\u751f\u6210\u56fe\u50cf\u3002<\/li>\n<\/ol>\n<h4>\u56fe\u50cf\u7f16\u8f91\uff08FLUX.1-Kontext-dev\uff09<\/h4>\n<ol>\n<li>\u52a0\u8f7d FLUX.1-Kontext-dev \u6a21\u578b\uff0c\u542f\u7528\u81ea\u7136\u8bed\u8a00\u56fe\u50cf\u7f16\u8f91\u529f\u80fd\u3002<\/li>\n<li>\u4e0a\u4f20\u9700\u8981\u7f16\u8f91\u7684\u56fe\u50cf\uff0c\u8f93\u5165\u7f16\u8f91\u6307\u4ee4\uff0c\u4f8b\u5982\u201cadd a bright moon in the sky\u201d\u3002<\/li>\n<li>\u8c03\u6574\u00a0<code>start_timestep<\/code>\u00a0\u548c\u00a0<code>end_timestep<\/code>\u00a0\u53c2\u6570\uff0c\u63a7\u5236\u7f16\u8f91\u7684\u5f3a\u5ea6\u548c\u8303\u56f4\u3002<\/li>\n<li>\u6267\u884c\u751f\u6210\uff0c\u67e5\u770b\u7f16\u8f91\u540e\u7684\u56fe\u50cf\u3002<\/li>\n<\/ol>\n<h4>\u8349\u56fe\u5230\u56fe\u50cf<\/h4>\n<ol>\n<li>\u5728 ComfyUI \u4e2d\u9009\u62e9\u00a0<code>Sketch-to-Image<\/code>\u00a0\u5de5\u4f5c\u6d41\u3002<\/li>\n<li>\u4e0a\u4f20\u8349\u56fe\u56fe\u50cf\uff0c\u8bbe\u7f6e\u751f\u6210\u53c2\u6570\u3002<\/li>\n<li>\u8fd0\u884c\u5de5\u4f5c\u6d41\uff0c\u751f\u6210\u57fa\u4e8e\u8349\u56fe\u7684\u56fe\u50cf\u3002<\/li>\n<\/ol>\n<h4>LoRA \u52a0\u8f7d<\/h4>\n<ol>\n<li>\u4e0b\u8f7d\u6216\u8bad\u7ec3\u81ea\u5b9a\u4e49 LoRA \u6a21\u578b\uff0c\u4fdd\u5b58\u5230\u00a0<code>models\/lora<\/code>\u00a0\u76ee\u5f55\u3002<\/li>\n<li>\u5728 ComfyUI \u5de5\u4f5c\u6d41\u4e2d\u6dfb\u52a0\u00a0<code>LoRA Loader<\/code>\u00a0\u8282\u70b9\uff0c\u6307\u5b9a LoRA \u6a21\u578b\u8def\u5f84\u3002<\/li>\n<li>\u8fd0\u884c\u5de5\u4f5c\u6d41\uff0c\u5e94\u7528 LoRA \u589e\u5f3a\u7684\u751f\u6210\u6548\u679c\u3002<\/li>\n<\/ol>\n<h3>\u6ce8\u610f\u4e8b\u9879<\/h3>\n<ul>\n<li>\u786e\u4fdd GPU \u67b6\u6784\u517c\u5bb9\uff08\u652f\u6301 Turing\u3001Ampere\u3001Ada\u3001A100 \u7b49\u67b6\u6784\uff09\u3002<\/li>\n<li>\u68c0\u67e5\u65e5\u5fd7\u8f93\u51fa\u4ee5\u786e\u8ba4 Python \u53ef\u6267\u884c\u6587\u4ef6\u8def\u5f84\uff0c\u7279\u522b\u662f\u5728 ComfyUI \u4fbf\u643a\u7248\u4e2d\u3002<\/li>\n<li>\u5982\u679c\u9047\u5230\u5b89\u88c5\u95ee\u9898\uff0c\u53ef\u53c2\u8003\u5b98\u65b9\u6587\u6863\u00a0<code>docs\/setup_windows.md<\/code>\u00a0\u6216\u6559\u7a0b\u89c6\u9891\u3002<\/li>\n<\/ul>\n<p>&nbsp;<\/p>\n<h2>\u5e94\u7528\u573a\u666f<\/h2>\n<ol>\n<li><strong>AI \u827a\u672f\u521b\u4f5c<\/strong><br \/>\n\u827a\u672f\u5bb6\u4f7f\u7528 Nunchaku \u751f\u6210\u9ad8\u8d28\u91cf\u7684\u827a\u672f\u56fe\u50cf\u3002\u7528\u6237\u8f93\u5165\u6587\u672c\u63cf\u8ff0\u6216\u4e0a\u4f20\u8349\u56fe\uff0c\u5feb\u901f\u751f\u6210\u7b26\u5408\u521b\u610f\u7684\u56fe\u50cf\uff0c\u9002\u5408\u6982\u5ff5\u8bbe\u8ba1\u548c\u827a\u672f\u63a2\u7d22\u3002<\/li>\n<li><strong>\u7814\u7a76\u4e0e\u5f00\u53d1<\/strong><br \/>\n\u7814\u7a76\u4eba\u5458\u5229\u7528 Nunchaku \u5728\u4f4e\u8d44\u6e90\u8bbe\u5907\u4e0a\u6d4b\u8bd5 4 \u4f4d\u91cf\u5316\u6a21\u578b\u3002\u5b83\u7684\u4f4e\u5185\u5b58\u5360\u7528\u548c\u9ad8\u63a8\u7406\u901f\u5ea6\u9002\u5408\u5b66\u672f\u5b9e\u9a8c\u548c\u6a21\u578b\u4f18\u5316\u3002<\/li>\n<li><strong>\u56fe\u50cf\u7f16\u8f91\u4e0e\u4fee\u590d<\/strong><br \/>\n\u6444\u5f71\u5e08\u6216\u8bbe\u8ba1\u5e08\u901a\u8fc7\u81ea\u7136\u8bed\u8a00\u6307\u4ee4\u7f16\u8f91\u56fe\u50cf\uff0c\u4f8b\u5982\u4fee\u590d\u7455\u75b5\u6216\u6dfb\u52a0\u5143\u7d20\uff0c\u7b80\u5316\u540e\u671f\u5904\u7406\u6d41\u7a0b\u3002<\/li>\n<li><strong>\u6559\u80b2\u4e0e\u5b66\u4e60<\/strong><br \/>\n\u5b66\u751f\u548c\u521d\u5b66\u8005\u4f7f\u7528 Nunchaku \u5b66\u4e60\u6269\u6563\u6a21\u578b\u7684\u63a8\u7406\u8fc7\u7a0b\u3002ComfyUI \u7684\u8282\u70b9\u5f0f\u754c\u9762\u76f4\u89c2\u6613\u61c2\uff0c\u9002\u5408\u6559\u5b66\u548c\u5b9e\u8df5\u3002<\/li>\n<\/ol>\n<p>&nbsp;<\/p>\n<h2>QA<\/h2>\n<ol>\n<li><strong>Nunchaku \u652f\u6301\u54ea\u4e9b GPU\uff1f<\/strong><br \/>\nNunchaku \u652f\u6301 NVIDIA GPU\uff0c\u5305\u62ec Turing\uff08RTX 2080\uff09\u3001Ampere\uff08RTX 3090, A6000\uff09\u3001Ada\uff08RTX 4090\uff09\u548c A100 \u67b6\u6784\u3002Blackwell GPU \u9700\u8981 CUDA 12.8 \u53ca PyTorch 2.7\u3002<\/li>\n<li><strong>\u5982\u4f55\u5728 Windows \u4e0a\u5b89\u88c5\uff1f<\/strong><br \/>\n\u5b89\u88c5 Visual Studio \u548c CUDA 12.6 \u4ee5\u4e0a\uff0c\u4f7f\u7528 Hugging Face \u7684 Windows wheel \u6587\u4ef6\u76f4\u63a5\u5b89\u88c5\uff0c\u6216\u53c2\u8003\u00a0<code>docs\/setup_windows.md<\/code>\u3002<\/li>\n<li><strong>4\u4f4d\u91cf\u5316\u4f1a\u5f71\u54cd\u56fe\u50cf\u8d28\u91cf\u5417\uff1f<\/strong><br \/>\nSVDQuant \u6280\u672f\u901a\u8fc7\u4f4e\u79e9\u7ec4\u4ef6\u5438\u6536\u5f02\u5e38\u503c\uff0c\u4fdd\u6301\u89c6\u89c9\u4fdd\u771f\u5ea6\u3002\u5728 FLUX.1-dev \u4e0a\uff0c4\u4f4d\u91cf\u5316\u7684\u56fe\u50cf\u8d28\u91cf\u63a5\u8fd1 16 \u4f4d\u6a21\u578b\u3002<\/li>\n<li><strong>\u5982\u4f55\u89e3\u51b3 ComfyUI \u5b89\u88c5\u5931\u8d25\uff1f<\/strong><br \/>\n\u786e\u4fdd Python \u73af\u5883\u4e0e ComfyUI \u4e00\u81f4\uff0c\u68c0\u67e5\u65e5\u5fd7\u4e2d\u7684 Python \u53ef\u6267\u884c\u6587\u4ef6\u8def\u5f84\u3002\u4f7f\u7528\u00a0<code>comfy-cli<\/code>\u00a0\u6216\u624b\u52a8\u5b89\u88c5\u00a0<code>ComfyUI-nunchaku<\/code>\u3002<\/li>\n<\/ol>\n<p>&nbsp;<\/p>\n<h2>\u5e94\u7528\u793a\u4f8b<\/h2>\n<p>nunchaku-flux.1-kontext-dev AI\u751f\u56fe\u4e00\u952e\u6574\u5408\u5305 \u6700\u4f4e4G \u663e\u5361\u53ef\u8fd0\u884c 30\u79d2\u51fa\u56fe v20250630<\/p>\n<p><strong>FLUX.1-<a href=\"https:\/\/www.kdjingpai.com\/tuichu-flux1-kontex\/\">Kontext<\/a>-dev \u6574\u5408\u5305<\/strong><\/p>\n<p><div class=\"huoduan_hide_box\" style=\"border:1px dashed #F60; padding:10px; margin:10px 0; line-height:200%; color:#F00; background-color:#FFF4FF; overflow:hidden; clear:both;\"><img loading=\"lazy\" decoding=\"async\" class=\"wxpic\" align=\"right\" src=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/05\/d8668ed8023fbe2.jpg\" style=\"width:150px;height:150px;margin-left:20px;display:inline;border:none\" width=\"150\" height=\"150\"  alt=\"AI\u751f\u4ea7\u529b\u5e94\u7528\" \/><span style=\"font-size:18px;\">\u6b64\u5904\u5185\u5bb9\u5df2\u7ecf\u88ab\u4f5c\u8005\u9690\u85cf\uff0c\u8bf7\u8f93\u5165\u9a8c\u8bc1\u7801\u67e5\u770b\u5185\u5bb9<\/span><form method=\"post\" style=\"margin:10px 0;\"><span class=\"yzts\" style=\"font-size:18px;float:left;\">\u9a8c\u8bc1\u7801\uff1a<\/span><input name=\"huoduan_verifycode\" id=\"verifycode\" type=\"text\" value=\"\" style=\"border:none;float:left;width:80px; height:32px; line-height:30px; padding:0 5px; border:1px solid #FF6600;-moz-border-radius: 0px;  -webkit-border-radius: 0px;  border-radius:0px;\" \/><input id=\"verifybtn\" style=\"border:none;float:left;width:80px; height:32px; line-height:32px; padding:0 5px; background-color:#F60; text-align:center; border:none; cursor:pointer; color:#FFF;-moz-border-radius: 0px; font-size:14px;  -webkit-border-radius: 0px;  border-radius:0px;\" name=\"\" type=\"submit\" value=\"\u63d0\u4ea4\u67e5\u770b\" \/><\/form><div style=\"clear:left;\"><\/div><span style=\"color:#00BF30\">\u8bf7\u5173\u6ce8\u672c\u7ad9\u5fae\u4fe1\u516c\u4f17\u53f7\uff0c\u56de\u590d\u201c<span style=\"color:blue\">\u9a8c\u8bc1\u7801<\/span>\u201d\uff0c\u83b7\u53d6\u9a8c\u8bc1\u7801\u3002\u5728\u5fae\u4fe1\u91cc\u641c\u7d22\u201c<span style=\"color:blue\">AI\u751f\u4ea7\u529b\u5e94\u7528<\/span>\u201d\u6216\u8005\u201c<span style=\"color:blue\">Artificial9527<\/span>\u201d\u6216\u8005\u5fae\u4fe1\u626b\u63cf\u53f3\u4fa7\u4e8c\u7ef4\u7801\u90fd\u53ef\u4ee5\u5173\u6ce8\u672c\u7ad9\u5fae\u4fe1\u516c\u4f17\u53f7\u3002<\/span><div class=\"cl\"><\/div><\/div><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Nunchaku \u662f\u7531 MIT HAN Lab \u5f00\u53d1\u7684\u4e00\u4e2a\u5f00\u6e90\u63a8\u7406\u5f15\u64ce\uff0c\u4e13\u6ce8\u4e8e\u9ad8\u6548\u8fd0\u884c4\u4f4d\u91cf\u5316\u7684\u6269\u6563\u6a21\u578b\u3002\u5b83\u901a\u8fc7 SVDQuant \u6280\u672f\uff0c\u5c06\u6a21\u578b\u7684\u6743\u91cd\u548c\u6fc0\u6d3b\u91cf\u5316\u4e3a4\u4f4d\uff0c\u663e\u8457\u964d\u4f4e\u5185\u5b58\u5360\u7528\u548c\u63a8\u7406\u5ef6\u8fdf\u3002Nunchaku \u652f\u6301 FLUX \u548c SAN&#8230;<\/p>\n","protected":false},"author":1,"featured_media":62357,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[230],"class_list":["post-31981","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-tool","tag-aikaiyuanxiangmu"],"_links":{"self":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/posts\/31981","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/comments?post=31981"}],"version-history":[{"count":0,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/posts\/31981\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/media\/62357"}],"wp:attachment":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/media?parent=31981"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/categories?post=31981"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/tags?post=31981"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}