{"id":20677,"date":"2026-04-29T21:23:44","date_gmt":"2026-04-29T13:23:44","guid":{"rendered":"https:\/\/92it.top\/?p=20677"},"modified":"2026-04-29T21:41:33","modified_gmt":"2026-04-29T13:41:33","slug":"mac-%e5%ae%89%e8%a3%85-llama-cpp-tips","status":"publish","type":"post","link":"https:\/\/92it.top\/?p=20677","title":{"rendered":"Mac \u5b89\u88c5 llama.cpp Tips"},"content":{"rendered":"\n<p><strong>\u524d\u8a00\ud83d\udd16<\/strong><\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p>llama.cpp\u662f\u4ee5\u4e00\u4e2a\u5f00\u6e90\u9879\u76ee\uff0c\u4e5f\u662f\u672c\u5730\u5316\u90e8\u7f72LLM\u6a21\u578b\u7684\u65b9\u5f0f\u4e4b\u4e00\uff0c\u9664\u4e86\u81ea\u8eab\u80fd\u591f\u4f5c\u4e3a\u5de5\u5177\u76f4\u63a5\u8fd0\u884c\u6a21\u578b\u6587\u4ef6\uff0c\u4e5f\u80fd\u591f\u88ab\u5176\u4ed6\u8f6f\u4ef6\u6216\u6846\u67b6\u8fdb\u884c\u8c03\u7528\u8fdb\u884c\u96c6\u6210\u3002<\/p>\n\n\n\n<p>llama.cpp \u662f\u7eafC++\u5b9e\u73b0\u7684\u5927\u6a21\u578b\u63a8\u7406\u6846\u67b6\uff0c\u6781\u81f4\u8f7b\u91cf\u5316\uff0c\u9002\u5408\u5bf9\u6027\u80fd\u6709\u6781\u81f4\u8981\u6c42\u7684\u573a\u666f\uff0c\u53ef\u76f4\u63a5\u8fd0\u884cGGUF\u683c\u5f0f\u7684\u91cf\u5316\u6a21\u578b\u3002<\/p>\n\n\n\n<p>github\u5730\u5740\uff1a<a href=\"https:\/\/github.com\/ggml-org\/llama.cpp\">https:\/\/github.com\/ggml-org\/llama.cpp<\/a><\/p>\n\n\n\n<p>llama.cpp \u662f\u7eaf C\/C++ \u5199\u7684\u5f00\u6e90 LLM \u63a8\u7406\u6846\u67b6\uff0c2023 \u5e74\u7531 Georgi Gerganov \u5f00\u53d1\uff0c\u6838\u5fc3\u76ee\u6807\u662f\u8ba9\u5927\u6a21\u578b\u5728\u666e\u901a\u7535\u8111\uff08CPU \/ \u82f9\u679c M \u82af\u7247\uff09\u4e0a\u9ad8\u6548\u8dd1\u8d77\u6765\uff0c\u51e0\u4e4e\u96f6\u5916\u90e8\u4f9d\u8d56\u3002<\/p>\n\n\n\n<p>\u6838\u5fc3\u7279\u70b9<\/p>\n\n\n\n<ul>\n<li><strong>\u6781\u81f4\u6027\u80fd<\/strong>\uff1a\u5e95\u5c42\u6c47\u7f16 \/ \u786c\u4ef6\u6307\u4ee4\u4f18\u5316\uff0c<strong>\u82f9\u679c M \u7cfb\u5217\uff08Metal\uff09\u4f18\u5316\u62c9\u6ee1<\/strong>\uff0c\u652f\u6301 GGUF \u91cf\u5316\uff08Q4_K_M \u7b49\uff09\uff0c\u5185\u5b58\u5360\u7528\u6781\u4f4e\u3002<\/li>\n\n\n\n<li><strong>\u9ad8\u5ea6\u7075\u6d3b<\/strong>\uff1a\u5168\u53c2\u6570\u53ef\u5b9a\u5236\uff08\u4e0a\u4e0b\u6587 <code>-c<\/code>\u3001GPU \u5c42 <code>-ngl<\/code>\u3001Flash Attention \u7b49\uff09\uff0c\u652f\u6301\u591a\u6a21\u6001\uff08mmproj\uff09\u3001\u81ea\u5b9a\u4e49\u505c\u6b62\u8bcd\u3001\u5173\u95ed\u601d\u8003\u8fc7\u7a0b\u3002<\/li>\n\n\n\n<li><strong>\u65e0\u4f9d\u8d56<\/strong>\uff1a\u7eaf C\/C++\uff0c\u7f16\u8bd1\u540e\u5355\u6587\u4ef6\uff0c\u53ef\u76f4\u63a5\u7528 <code>llama-server<\/code> \u542f API \/ \u7f51\u9875\u670d\u52a1\u3002<\/li>\n\n\n\n<li><strong>\u4f60\u7684\u7528\u6cd5<\/strong>\uff1a\u4f60\u73b0\u5728\u7528\u7684\u547d\u4ee4\uff08<code>.\/llama-server -m ...<\/code>\uff09\u5c31\u662f\u76f4\u63a5\u8c03\u7528\u539f\u751f llama.cpp\uff0c<strong>\u6027\u80fd\u6700\u5f3a\u3001\u53ef\u63a7\u6027\u6700\u9ad8<\/strong>\u3002<\/li>\n<\/ul>\n\n\n\n<p>\u3000\u3000<\/p>\n\n\n\n<p><strong>Mac \u5b89\u88c5 llama.cpp\ud83d\udd16<\/strong><\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p><strong>\ud83d\udd391. \u4e0b\u8f7dllama\u4ee3\u7801<\/strong><\/p>\n\n\n\n<pre class=\"EnlighterJSRAW\" data-enlighter-language=\"generic\" data-enlighter-theme=\"\" data-enlighter-highlight=\"\" data-enlighter-linenumbers=\"\" data-enlighter-lineoffset=\"\" data-enlighter-title=\"\" data-enlighter-group=\"\">git clone https:\/\/github.com\/ggerganov\/llama.cpp.git<\/pre>\n\n\n\n<p>\u3000\u3000<\/p>\n\n\n\n<p><strong>\ud83d\udd392.\u7f16\u8bd1llama.cpp<\/strong><\/p>\n\n\n\n<p>\u5b98\u7f51build\u8bf4\u660e\uff1a<a href=\"https:\/\/github.com\/ggml-org\/llama.cpp\/blob\/master\/docs\/build.md#metal-build\">https:\/\/github.com\/ggml-org\/llama.cpp\/blob\/master\/docs\/build.md#metal-build<\/a><\/p>\n\n\n\n<pre class=\"EnlighterJSRAW\" data-enlighter-language=\"generic\" data-enlighter-theme=\"\" data-enlighter-highlight=\"\" data-enlighter-linenumbers=\"\" data-enlighter-lineoffset=\"\" data-enlighter-title=\"\" data-enlighter-group=\"\">cd llama.cpp  \/\/ cd\u5230\u5de5\u7a0b\u76ee\u5f55\ncmake -B build \/\/ \u51c6\u5907\u7f16\u8bd1\u73af\u5883\ncmake --build build --config Release \/\/ \u6b63\u5f0f\u7f16\u8bd1\u51fa\u53ef\u6267\u884c\u6587\u4ef6<\/pre>\n\n\n<div class=\"wp-block-image\">\n<figure class=\"aligncenter size-large is-resized\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"708\" src=\"https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-165-1024x708.png\" alt=\"\" class=\"wp-image-20679\" style=\"width:452px;height:auto\" srcset=\"https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-165-1024x708.png 1024w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-165-300x208.png 300w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-165-768x531.png 768w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-165-1536x1062.png 1536w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-165-830x574.png 830w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-165-230x159.png 230w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-165-350x242.png 350w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-165-480x332.png 480w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-165.png 1680w\" sizes=\"(max-width: 1024px) 100vw, 1024px\" \/><\/figure><\/div>\n\n\n<p>\u56e0\u4e3a\u7f16\u8bd1\u9700\u8981cmake\uff0c\u5982\u679c\u51fa\u73b0\u4e0b\u9762\u7684\u9519\u8bef\uff0c\u8bf7\u5148\u5b89\u88c5cmake<\/p>\n\n\n\n<pre class=\"EnlighterJSRAW\" data-enlighter-language=\"generic\" data-enlighter-theme=\"\" data-enlighter-highlight=\"\" data-enlighter-linenumbers=\"\" data-enlighter-lineoffset=\"\" data-enlighter-title=\"\" data-enlighter-group=\"\">MacBook-Air llama.cpp % make\nMakefile:6: *** Build system changed:\n The Makefile build has been replaced by CMake.\n\n For build instructions see:\n https:\/\/github.com\/ggml-org\/llama.cpp\/blob\/master\/docs\/build.md\n\n.  Stop.<\/pre>\n\n\n\n<p>\u53ef\u4ee5\u7528\u4e0b\u9762\u7684\u547d\u4ee4\u5b89\u88c5 cmake<\/p>\n\n\n\n<pre class=\"EnlighterJSRAW\" data-enlighter-language=\"generic\" data-enlighter-theme=\"\" data-enlighter-highlight=\"\" data-enlighter-linenumbers=\"\" data-enlighter-lineoffset=\"\" data-enlighter-title=\"\" data-enlighter-group=\"\">brew install cmake<\/pre>\n\n\n\n<p>\u3000\u3000<\/p>\n\n\n\n<p><strong>\ud83d\udd393.\u4e0b\u8f7dgguf\u6a21\u578b<\/strong><\/p>\n\n\n\n<p>\u7136\u540e\u6211\u4eec\u5c31\u53ef\u4ee5\u6109\u5feb\u7684\u53bb huggingface \u7f51\u7ad9\u4e0b\u8f7dgguf LLM\u6a21\u578b\u4e86\u3002\u6bd4\u5982\u4e0b\u9762\u7684\u662fqwen3.5\u6a21\u578b<\/p>\n\n\n\n<p><a href=\"https:\/\/huggingface.co\/HauhauCS\/Qwen3.5-4B-Uncensored-HauhauCS-Aggressive\/tree\/main\">https:\/\/huggingface.co\/HauhauCS\/Qwen3.5-4B-Uncensored-HauhauCS-Aggressive\/tree\/main<\/a> <\/p>\n\n\n<div class=\"wp-block-image\">\n<figure class=\"aligncenter size-large is-resized\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"651\" src=\"https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-167-1024x651.png\" alt=\"\" class=\"wp-image-20681\" style=\"width:460px;height:auto\" srcset=\"https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-167-1024x651.png 1024w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-167-300x191.png 300w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-167-768x488.png 768w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-167-1536x977.png 1536w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-167-2048x1302.png 2048w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-167-830x528.png 830w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-167-230x146.png 230w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-167-350x223.png 350w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-167-480x305.png 480w\" sizes=\"(max-width: 1024px) 100vw, 1024px\" \/><\/figure><\/div>\n\n\n<p>\u56e0\u4e3a\u662f\u591a\u6a21\u6001\u6a21\u578b\uff0c\u6709mmproj\u6587\u4ef6\u3002<\/p>\n\n\n\n<p>\u6211\u4eec\u53ef\u4ee5\u628a\u4e0b\u8f7d\u7684\u6a21\u578b\u90fd\u653e\u5728\u4e0b\u9762\u7684\u8def\u5f84\u4e2d<\/p>\n\n\n\n<pre class=\"EnlighterJSRAW\" data-enlighter-language=\"generic\" data-enlighter-theme=\"\" data-enlighter-highlight=\"\" data-enlighter-linenumbers=\"\" data-enlighter-lineoffset=\"\" data-enlighter-title=\"\" data-enlighter-group=\"\">llama\/llama.cpp\/models\/<\/pre>\n\n\n<div class=\"wp-block-image\">\n<figure class=\"aligncenter size-large is-resized\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"411\" src=\"https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-170-1024x411.png\" alt=\"\" class=\"wp-image-20684\" style=\"width:414px;height:auto\" srcset=\"https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-170-1024x411.png 1024w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-170-300x120.png 300w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-170-768x308.png 768w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-170-830x333.png 830w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-170-230x92.png 230w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-170-350x140.png 350w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-170-480x193.png 480w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-170.png 1356w\" sizes=\"(max-width: 1024px) 100vw, 1024px\" \/><\/figure><\/div>\n\n\n<p>\u3000\u3000<\/p>\n\n\n\n<p><strong>\u542f\u52a8 llama.cpp\ud83d\udd16<\/strong><\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p><strong>\ud83d\udd39llama.cpp \u53ef\u4ee5\u4ee5web server\u7684\u6a21\u5f0f\u542f\u52a8<\/strong><\/p>\n\n\n\n<pre class=\"EnlighterJSRAW\" data-enlighter-language=\"generic\" data-enlighter-theme=\"\" data-enlighter-highlight=\"\" data-enlighter-linenumbers=\"\" data-enlighter-lineoffset=\"\" data-enlighter-title=\"\" data-enlighter-group=\"\">cd llama\/llama.cpp\/build\/bin    \\\\\u5207\u6362\u5230llama bin\u76ee\u5f55\u4e0b\n\n.\/llama-server \\\n  -m ..\/..\/models\/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-Q4_K_M.gguf \\\n  --mmproj ..\/..\/models\/mmproj-Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-BF16.gguf \\\n  --port 8080 \\\n  -c 16384 \\\n  --n-gpu-layers 35 \\\n  --chat-template-kwargs '{\"enable_thinking\":false}'\n  --api-key \"MySecret123\"<\/pre>\n\n\n\n<p>\u6d4f\u89c8\u5668\u6253\u5f00 <a href=\"http:\/\/127.0.0.1:8080\/\">http:\/\/127.0.0.1:8080\/<\/a> \u7f51\u9875\u4f1a\u63d0\u793a\u8f93\u5165API Key<\/p>\n\n\n<div class=\"wp-block-image\">\n<figure class=\"aligncenter size-large is-resized\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"720\" src=\"https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-171-1024x720.png\" alt=\"\" class=\"wp-image-20685\" style=\"width:442px;height:auto\" srcset=\"https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-171-1024x720.png 1024w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-171-300x211.png 300w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-171-768x540.png 768w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-171-1536x1080.png 1536w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-171-830x584.png 830w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-171-230x162.png 230w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-171-350x246.png 350w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-171-480x338.png 480w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-171.png 1638w\" sizes=\"(max-width: 1024px) 100vw, 1024px\" \/><\/figure><\/div>\n\n\n<p>\u8f93\u5165\u6b63\u786e\u7684key\u4ee5\u540e\uff0c\u5c31\u53ef\u4ee5\u6109\u5feb\u7684\u804a\u5929\u4e86<\/p>\n\n\n<div class=\"wp-block-image\">\n<figure class=\"aligncenter size-large is-resized\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"651\" src=\"https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-172-1024x651.png\" alt=\"\" class=\"wp-image-20686\" style=\"width:450px;height:auto\" srcset=\"https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-172-1024x651.png 1024w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-172-300x191.png 300w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-172-768x488.png 768w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-172-1536x977.png 1536w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-172-830x528.png 830w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-172-230x146.png 230w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-172-350x223.png 350w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-172-480x305.png 480w, https:\/\/92it.top\/wp-content\/uploads\/2026\/04\/\u56fe\u7247-172.png 2048w\" sizes=\"(max-width: 1024px) 100vw, 1024px\" \/><\/figure><\/div>","protected":false},"excerpt":{"rendered":"<p>\u524d\u8a00\ud83d\udd16 llama.cpp\u662f\u4ee5\u4e00\u4e2a\u5f00\u6e90\u9879\u76ee\uff0c\u4e5f\u662f\u672c\u5730\u5316\u90e8\u7f72LLM\u6a21\u578b\u7684\u65b9\u5f0f\u4e4b\u4e00\uff0c\u9664\u4e86\u81ea\u8eab\u80fd\u591f\u4f5c\u4e3a\u5de5\u5177\u76f4\u63a5\u8fd0\u884c\u6a21 [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[48],"tags":[],"_links":{"self":[{"href":"https:\/\/92it.top\/index.php?rest_route=\/wp\/v2\/posts\/20677"}],"collection":[{"href":"https:\/\/92it.top\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/92it.top\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/92it.top\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/92it.top\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=20677"}],"version-history":[{"count":2,"href":"https:\/\/92it.top\/index.php?rest_route=\/wp\/v2\/posts\/20677\/revisions"}],"predecessor-version":[{"id":20688,"href":"https:\/\/92it.top\/index.php?rest_route=\/wp\/v2\/posts\/20677\/revisions\/20688"}],"wp:attachment":[{"href":"https:\/\/92it.top\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=20677"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/92it.top\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=20677"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/92it.top\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=20677"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}