{"id":13647,"date":"2024-11-20T13:09:15","date_gmt":"2024-11-20T05:09:15","guid":{"rendered":"https:\/\/www.aisharenet.com\/?p=13647"},"modified":"2025-02-17T15:18:34","modified_gmt":"2025-02-17T07:18:34","slug":"llamacpp","status":"publish","type":"post","link":"https:\/\/www.kdjingpai.com\/de\/llamacpp\/","title":{"rendered":"llama.cpp\uff1a\u9ad8\u6548\u63a8\u7406\u5de5\u5177\uff0c\u652f\u6301\u591a\u79cd\u786c\u4ef6\uff0c\u8f7b\u677e\u5b9e\u73b0LLM\u63a8\u7406"},"content":{"rendered":"<p>llama.cpp\u662f\u4e00\u4e2a\u7528\u7eafC\/C++\u5b9e\u73b0\u7684\u5e93\uff0c\u65e8\u5728\u7b80\u5316\u5927\u8bed\u8a00\u6a21\u578b\uff08LLM\uff09\u7684\u63a8\u7406\u8fc7\u7a0b\u3002\u5b83\u652f\u6301\u591a\u79cd\u786c\u4ef6\u5e73\u53f0\uff0c\u5305\u62ecApple Silicon\u3001NVIDIA GPU\u548cAMD GPU\uff0c\u5e76\u63d0\u4f9b\u591a\u79cd\u91cf\u5316\u9009\u9879\u4ee5\u63d0\u9ad8\u63a8\u7406\u901f\u5ea6\u548c\u51cf\u5c11\u5185\u5b58\u4f7f\u7528\u3002\u8be5\u9879\u76ee\u7684\u76ee\u6807\u662f\u901a\u8fc7\u6700\u5c11\u7684\u8bbe\u7f6e\u5b9e\u73b0\u9ad8\u6027\u80fd\u7684LLM\u63a8\u7406\uff0c\u9002\u7528\u4e8e\u672c\u5730\u548c\u4e91\u7aef\u73af\u5883\u3002<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter  wp-image-21420\" title=\"llama.cpp\uff1a\u9ad8\u6548\u63a8\u7406\u5de5\u5177\uff0c\u652f\u6301\u591a\u79cd\u786c\u4ef6\uff0c\u8f7b\u677e\u5b9e\u73b0LLM\u63a8\u7406-1\" src=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/02\/5d4e9f1bd340110.jpg\" alt=\"llama.cpp\uff1a\u9ad8\u6548\u63a8\u7406\u5de5\u5177\uff0c\u652f\u6301\u591a\u79cd\u786c\u4ef6\uff0c\u8f7b\u677e\u5b9e\u73b0LLM\u63a8\u7406-1\" width=\"1029\" height=\"343\" srcset=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/02\/5d4e9f1bd340110.jpg 1500w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/02\/5d4e9f1bd340110-768x256.jpg 768w\" sizes=\"auto, (max-width: 1029px) 100vw, 1029px\" \/><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-13648\" title=\"llama.cpp\uff1a\u9ad8\u6548\u63a8\u7406\u5de5\u5177\uff0c\u652f\u6301\u591a\u79cd\u786c\u4ef6\uff0c\u8f7b\u677e\u5b9e\u73b0LLM\u63a8\u7406-1\" src=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2024\/11\/39d7b79382e27a9.jpg\" alt=\"llama.cpp\uff1a\u9ad8\u6548\u63a8\u7406\u5de5\u5177\uff0c\u652f\u6301\u591a\u79cd\u786c\u4ef6\uff0c\u8f7b\u677e\u5b9e\u73b0LLM\u63a8\u7406-1\" width=\"1010\" height=\"567\" srcset=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2024\/11\/39d7b79382e27a9.jpg 1010w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2024\/11\/39d7b79382e27a9-300x168.jpg 300w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2024\/11\/39d7b79382e27a9-768x431.jpg 768w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2024\/11\/39d7b79382e27a9-18x10.jpg 18w\" sizes=\"auto, (max-width: 1010px) 100vw, 1010px\" \/><\/p>\n<p>&nbsp;<\/p>\n<h2>\u529f\u80fd\u5217\u8868<\/h2>\n<ul>\n<li>\u652f\u6301\u591a\u79cd\u786c\u4ef6\u5e73\u53f0\uff0c\u5305\u62ecApple Silicon\u3001NVIDIA GPU\u548cAMD GPU<\/li>\n<li>\u63d0\u4f9b1.5\u4f4d\u52308\u4f4d\u7684\u6574\u6570\u91cf\u5316\u9009\u9879<\/li>\n<li>\u652f\u6301\u591a\u79cdLLM\u6a21\u578b\uff0c\u5982LLaMA\u3001Mistral\u3001Falcon\u7b49<\/li>\n<li>\u63d0\u4f9bREST API\u63a5\u53e3\uff0c\u4fbf\u4e8e\u96c6\u6210<\/li>\n<li>\u652f\u6301CPU+GPU\u6df7\u5408\u63a8\u7406<\/li>\n<li>\u63d0\u4f9b\u591a\u79cd\u7f16\u7a0b\u8bed\u8a00\u7ed1\u5b9a\uff0c\u5982Python\u3001Go\u3001Node.js\u7b49<\/li>\n<li>\u63d0\u4f9b\u591a\u79cd\u5de5\u5177\u548c\u57fa\u7840\u8bbe\u65bd\u652f\u6301\uff0c\u5982\u6a21\u578b\u8f6c\u6362\u5de5\u5177\u3001\u8d1f\u8f7d\u5747\u8861\u5668\u7b49<\/li>\n<\/ul>\n<p>&nbsp;<\/p>\n<h2>\u4f7f\u7528\u5e2e\u52a9<\/h2>\n<h3>\u5b89\u88c5\u6d41\u7a0b<\/h3>\n<ol>\n<li>\u514b\u9686\u4ed3\u5e93\uff1a<\/li>\n<\/ol>\n<pre><code>   git clone https:\/\/github.com\/ggerganov\/llama.cpp.git\r\ncd llama.cpp\r\n<\/code><\/pre>\n<ol start=\"2\">\n<li>\u7f16\u8bd1\u9879\u76ee\uff1a<\/li>\n<\/ol>\n<pre><code>   make\r\n<\/code><\/pre>\n<h3>\u4f7f\u7528\u6307\u5357<\/h3>\n<h4>\u6a21\u578b\u8f6c\u6362<\/h4>\n<p>llama.cpp\u63d0\u4f9b\u4e86\u591a\u79cd\u5de5\u5177\u6765\u8f6c\u6362\u548c\u91cf\u5316\u6a21\u578b\uff0c\u4ee5\u4fbf\u5728\u4e0d\u540c\u786c\u4ef6\u4e0a\u9ad8\u6548\u8fd0\u884c\u3002\u4f8b\u5982\uff0c\u53ef\u4ee5\u4f7f\u7528\u4ee5\u4e0b\u547d\u4ee4\u5c06Hugging Face\u6a21\u578b\u8f6c\u6362\u4e3aGGML\u683c\u5f0f\uff1a<\/p>\n<pre><code>python3 convert_hf_to_gguf.py --model &lt;model_name&gt;\r\n<\/code><\/pre>\n<h4>\u63a8\u7406\u793a\u4f8b<\/h4>\n<p>\u7f16\u8bd1\u5b8c\u6210\u540e\uff0c\u53ef\u4ee5\u4f7f\u7528\u4ee5\u4e0b\u547d\u4ee4\u8fdb\u884c\u63a8\u7406\uff1a<\/p>\n<pre><code>.\/llama-cli -m models\/llama-13b-v2\/ggml-model-q4_0.gguf -p \"\u4f60\u597d\uff0c\u4e16\u754c\uff01\"\r\n<\/code><\/pre>\n<h4>REST API\u4f7f\u7528<\/h4>\n<p>llama.cpp\u8fd8\u63d0\u4f9b\u4e86\u4e00\u4e2a\u517c\u5bb9OpenAI API\u7684HTTP\u670d\u52a1\u5668\uff0c\u53ef\u4ee5\u7528\u4e8e\u672c\u5730\u6a21\u578b\u7684\u63a8\u7406\u670d\u52a1\u3002\u542f\u52a8\u670d\u52a1\u5668\uff1a<\/p>\n<pre><code>.\/llama-server -m models\/llama-13b-v2\/ggml-model-q4_0.gguf --port 8080\r\n<\/code><\/pre>\n<p>\u7136\u540e\u53ef\u4ee5\u901a\u8fc7\u6d4f\u89c8\u5668\u8bbf\u95ee\u57fa\u672c\u7684Web UI\u6216\u4f7f\u7528API\u8fdb\u884c\u63a8\u7406\u8bf7\u6c42\uff1a<\/p>\n<pre><code>curl -X POST http:\/\/localhost:8080\/v1\/chat -d '{\"prompt\": \"\u4f60\u597d\uff0c\u4e16\u754c\uff01\"}'\r\n<\/code><\/pre>\n<h3>\u8be6\u7ec6\u529f\u80fd\u64cd\u4f5c\u6d41\u7a0b<\/h3>\n<ol>\n<li><strong>\u6a21\u578b\u52a0\u8f7d<\/strong>\uff1a\u9996\u5148\u9700\u8981\u5c06\u6a21\u578b\u6587\u4ef6\u4e0b\u8f7d\u5e76\u653e\u7f6e\u5728\u6307\u5b9a\u76ee\u5f55\uff0c\u7136\u540e\u4f7f\u7528\u547d\u4ee4\u884c\u5de5\u5177\u52a0\u8f7d\u6a21\u578b\u3002<\/li>\n<li><strong>\u63a8\u7406\u914d\u7f6e<\/strong>\uff1a\u53ef\u4ee5\u901a\u8fc7\u914d\u7f6e\u6587\u4ef6\u6216\u547d\u4ee4\u884c\u53c2\u6570\u8bbe\u7f6e\u63a8\u7406\u7684\u76f8\u5173\u53c2\u6570\uff0c\u5982\u4e0a\u4e0b\u6587\u957f\u5ea6\u3001\u6279\u5904\u7406\u5927\u5c0f\u7b49\u3002<\/li>\n<li><strong>API\u96c6\u6210<\/strong>\uff1a\u901a\u8fc7REST API\u63a5\u53e3\uff0c\u53ef\u4ee5\u5c06llama.cpp\u96c6\u6210\u5230\u73b0\u6709\u7684\u5e94\u7528\u7a0b\u5e8f\u4e2d\uff0c\u5b9e\u73b0\u81ea\u52a8\u5316\u63a8\u7406\u670d\u52a1\u3002<\/li>\n<li><strong>\u6027\u80fd\u4f18\u5316<\/strong>\uff1a\u5229\u7528\u91cf\u5316\u9009\u9879\u548c\u786c\u4ef6\u52a0\u901f\u529f\u80fd\uff0c\u53ef\u4ee5\u663e\u8457\u63d0\u9ad8\u63a8\u7406\u901f\u5ea6\u548c\u6548\u7387\u3002<\/li>\n<\/ol>\n","protected":false},"excerpt":{"rendered":"<p>llama.cpp\u662f\u4e00\u4e2a\u7528\u7eafC\/C++\u5b9e\u73b0\u7684\u5e93\uff0c\u65e8\u5728\u7b80\u5316\u5927\u8bed\u8a00\u6a21\u578b\uff08LLM\uff09\u7684\u63a8\u7406\u8fc7\u7a0b\u3002\u5b83\u652f\u6301\u591a\u79cd\u786c\u4ef6\u5e73\u53f0\uff0c\u5305\u62ecApple Silicon\u3001NVIDIA GPU\u548cAMD GPU\uff0c\u5e76\u63d0\u4f9b\u591a\u79cd\u91cf\u5316\u9009\u9879\u4ee5\u63d0\u9ad8\u63a8\u7406\u901f\u5ea6\u548c\u51cf\u5c11\u5185\u5b58\u4f7f\u7528\u3002\u8be5\u9879\u76ee\u7684\u76ee\u6807\u662f&#8230;<\/p>\n","protected":false},"author":1,"featured_media":61270,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[232],"class_list":["post-13647","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-tool","tag-bendebushukaiyuanba"],"_links":{"self":[{"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/posts\/13647","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/comments?post=13647"}],"version-history":[{"count":0,"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/posts\/13647\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/media\/61270"}],"wp:attachment":[{"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/media?parent=13647"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/categories?post=13647"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/tags?post=13647"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}