{"id":89586,"date":"2026-08-03T09:57:25","date_gmt":"2026-08-03T01:57:25","guid":{"rendered":"https:\/\/www.wsisp.com\/helps\/89586.html"},"modified":"2026-08-03T09:57:25","modified_gmt":"2026-08-03T01:57:25","slug":"vllm-tensorrt-llm-sglang-%e6%a8%aa%e8%af%84%ef%bc%9a%e6%88%91%e5%9c%a8-windows-%e4%b8%8a%e8%bf%9e-pip-install-%e9%83%bd%e8%a3%85%e4%b8%8d%e4%b8%8a","status":"publish","type":"post","link":"https:\/\/www.wsisp.com\/helps\/89586.html","title":{"rendered":"vLLM \/ TensorRT-LLM \/ SGLang \u6a2a\u8bc4\uff1a\u6211\u5728 Windows \u4e0a\u8fde pip install \u90fd\u88c5\u4e0d\u4e0a"},"content":{"rendered":"<h2 id=\"vllm-tensorrt-llm-sglang-windows-pip-install\"><\/h2>\n<p>\u7cfb\u5217\u7b2c 1 \u7bc7 \u00b7 \u63a8\u7406\u9009\u578b\u51b3\u7b56<\/p>\n<p>\u5b98\u7f51\u8bf4 vLLM \u6bd4 HuggingFace Transformers \u5feb 3 \u500d&#xff0c;\u4f46\u6211\u5728\u81ea\u5df1\u7684 RTX 5060 \u7b14\u8bb0\u672c\u4e0a\u6572 pip install vllm&#xff0c;\u76f4\u63a5\u62a5\u300c\u627e\u4e0d\u5230\u5339\u914d\u7684 Windows \u5305\u300d\u3002\u5148\u522b\u76f2\u4fe1 benchmark&#xff0c;\u90e8\u7f72\u95e8\u69db\u624d\u662f\u7b2c\u4e00\u9053\u5173\u3002<\/p>\n<p>\u672c\u8f91\u4e3b\u7ebf&#xff1a;\u628a\u5927\u6a21\u578b\u4ece\u300c\u80fd\u8c03\u901a\u300d\u505a\u5230\u300c\u8dd1\u5f97\u5feb\u3001\u8dd1\u5f97\u7701\u3001\u6d4b\u5f97\u51c6\u3001\u5b88\u5f97\u4f4f\u300d\u3002\u8fd9\u4e00\u7bc7\u5148\u89e3\u51b3\u9009\u578b\u2014\u2014\u540c\u4e00\u6a21\u578b\u3001\u540c\u4e00\u786c\u4ef6&#xff0c;\u4e09\u4e2a\u63a8\u7406\u5f15\u64ce\u5230\u5e95\u600e\u4e48\u9009\u3002<\/p>\n<hr \/>\n<h3 id=\"api\">\u4e00\u3001\u4e3a\u4ec0\u4e48\u4e0d\u76f4\u63a5\u7528 API&#xff1f;<\/h3>\n<p>API \u591f\u7528&#xff0c;\u4f46\u4e09\u4e2a\u52a8\u56e0\u8fdf\u65e9\u628a\u4f60\u63a8\u5411\u81ea\u5efa\u63a8\u7406\u5f15\u64ce&#xff1a;<\/p>\n<ul>\n<li>\u5e76\u53d1&#xff1a;API \u6309 token \u8ba1\u8d39&#xff0c;\u9ad8\u5e76\u53d1\u4e0b\u6210\u672c\u5931\u63a7&#xff0c;\u4e14\u901f\u7387\u9650\u5236\u5361\u8116\u5b50\u3002<\/li>\n<li>\u6210\u672c&#xff1a;\u957f\u6587\u672c&#043;\u9ad8 QPS \u573a\u666f&#xff0c;\u81ea\u6258\u7ba1\u4e00\u5f20\u5361\u5f80\u5f80\u6bd4 API \u8d26\u5355\u4fbf\u5b9c\u4e00\u4e2a\u6570\u91cf\u7ea7\u3002<\/li>\n<li>\u79c1\u6709\u5316&#xff1a;\u6570\u636e\u4e0d\u51fa\u5185\u7f51&#xff0c;\u91d1\u878d\/\u533b\u7597\/\u653f\u4f01\u7684\u786c\u8981\u6c42\u3002<\/li>\n<\/ul>\n<p>\u4f46\u81ea\u5efa\u7684\u7b2c\u4e00\u8bfe\u4e0d\u662f\u300c\u8c01\u6700\u5feb\u300d&#xff0c;\u662f\u300c\u8c01\u88c5\u5f97\u4e0a\u300d\u3002<\/p>\n<hr \/>\n<h3 id=\"_1\">\u4e8c\u3001\u4e09\u6846\u67b6\u5b9a\u4f4d\u5dee\u5f02&#xff08;\u4e00\u773c\u770b\u61c2&#xff09;<\/h3>\n<table>\n<tr>\u5f15\u64ce\u5b9a\u4f4d\u5f3a\u9879\u4ee3\u4ef7<\/tr>\n<tbody>\n<tr>\n<td>vLLM<\/td>\n<td>\u901a\u7528\u5747\u8861<\/td>\n<td>\u793e\u533a\u5927\u3001\u8fde\u7eed\u6279\u5904\u7406\u6210\u719f\u3001\u6613\u4e0a\u624b<\/td>\n<td>\u6781\u81f4\u6027\u80fd\u975e\u6700\u5f3a<\/td>\n<\/tr>\n<tr>\n<td>SGLang<\/td>\n<td>\u7ed3\u6784\u5316\u8f93\u51fa<\/td>\n<td>RadixAttention \u590d\u7528\u524d\u7f00\u3001Agent\/\u591a\u8f6e\u5f3a<\/td>\n<td>\u914d\u7f6e\u5fc3\u667a\u8d1f\u62c5\u5927<\/td>\n<\/tr>\n<tr>\n<td>TensorRT-LLM<\/td>\n<td>N \u5361\u6781\u81f4<\/td>\n<td>NVIDIA \u4e0a\u5ef6\u8fdf\/\u541e\u5410\u6700\u4f18<\/td>\n<td>\u7f16\u8bd1\u91cd\u3001\u7248\u672c\u9501\u6b7b\u3001Linux only<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<hr \/>\n<h3 id=\"_2\">\u4e09\u3001\u5b9e\u6d4b\u73af\u5883&#xff08;\u672c\u6b21\u771f\u8dd1\u7684\u90e8\u5206&#xff09;<\/h3>\n<ul>\n<li>GPU&#xff1a;NVIDIA RTX 5060 Laptop GPU&#xff0c;8GB \u663e\u5b58&#xff0c;Blackwell \u67b6\u6784&#xff08;\u8ba1\u7b97\u80fd\u529b 12.0&#xff09;<\/li>\n<li>\u9a71\u52a8&#xff1a;610.62<\/li>\n<li>\u7cfb\u7edf&#xff1a;Windows 11&#xff08;win32&#xff09;<\/li>\n<li>\u6846\u67b6&#xff1a;Python 3.13&#xff0c;torch&#xff08;CUDA 12.8 \u7248&#xff09;\u5df2\u5b89\u88c5\u5e76\u9a8c\u8bc1 CUDA \u77e9\u9635\u4e58\u53ef\u7528<\/li>\n<li>\u6570\u636e\u8bf4\u660e&#xff1a;\u2460\u300c\u5b89\u88c5\u53ef\u88c5\u6027\u300d\u4e3a\u672c\u4eba\u5728\u672c\u673a\/PyPI \u5b9e\u8dd1\u6838\u9a8c&#xff1b;\u2461\u300c\u6027\u80fd\u541e\u5410\u300d\u4e3a\u5404\u9879\u76ee\u5b98\u65b9 benchmark \u516c\u5f00\u6570\u636e&#xff08;\u975e\u672c\u4eba\u5b9e\u6d4b&#xff09;&#xff0c;\u6587\u672b\u9644\u6765\u6e90&#xff1b;\u2462\u5199\u4f5c\u65f6 ModelScope \u6a21\u578b\u6e90\u4e0d\u53ef\u8fbe&#xff0c;\u672c\u5730 Qwen2.5 \u63a8\u7406\u57fa\u51c6\u63a8\u8fdf\u5230\u540e\u7eed\u8865\u6d4b\u3002<\/li>\n<\/ul>\n<hr \/>\n<h3 id=\"windows\">\u56db\u3001\u7b2c\u4e00\u5173&#xff1a;Windows \u5230\u5e95\u80fd\u4e0d\u80fd\u88c5&#xff1f;&#xff08;\u672c\u4eba\u5b9e\u8dd1&#xff09;<\/h3>\n<p>\u6211\u7528 PyPI JSON API \u62c9\u4e86\u4e09\u8005\u6700\u65b0\u7248\u7684 wheel \u5e73\u53f0\u6807\u7b7e&#xff0c;\u7ed3\u8bba\u5f88\u76f4\u63a5&#xff1a;<\/p>\n<table>\n<tr>\u5f15\u64ce(\u6700\u65b0\u7248)PyPI wheel \u5e73\u53f0Windows wheelCUDA wheel\u672c\u673a pip install<\/tr>\n<tbody>\n<tr>\n<td>vLLM 0.26.0<\/td>\n<td>manylinux_2_28 (x86\/aarch64)<\/td>\n<td>\u65e0<\/td>\n<td>\u65e0<\/td>\n<td>\u274c \u65e0\u5339\u914d\u5305<\/td>\n<\/tr>\n<tr>\n<td>SGLang 0.5.16<\/td>\n<td>manylinux_2_34 (x86\/aarch64)<\/td>\n<td>\u65e0<\/td>\n<td>\u65e0<\/td>\n<td>\u274c \u65e0\u5339\u914d\u5305<\/td>\n<\/tr>\n<tr>\n<td>TensorRT-LLM 1.2.1<\/td>\n<td>\u65e0\u4efb\u4f55 wheel&#xff08;\u4ec5\u6e90\u7801\u5305&#xff09;<\/td>\n<td>\u65e0<\/td>\n<td>\u65e0<\/td>\n<td>\u274c \u9700\u6e90\u7801\u7f16\u8bd1<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u53cd\u76f4\u89c9\u7ed3\u8bba&#xff1a;\u4e09\u5927\u5f15\u64ce\u6700\u65b0\u7248\u5728 PyPI \u4e0a\u5168\u90fd\u53ea\u53d1 Linux wheel&#xff0c;\u6ca1\u6709 Windows \u8f6e\u5b50\u3002\u60f3\u5728 Windows \u6d88\u8d39\u7ea7\u663e\u5361\u4e0a\u76f4\u63a5 pip install&#xff0c;\u76ee\u524d\u4e09\u4e2a\u90fd\u88c5\u4e0d\u4e0a\u3002<\/p>\n<p>\u8840\u7684\u6559\u8bad&#xff1a;\u522b\u4fe1\u300c\u67d0\u67d0\u6846\u67b6\u652f\u6301 Windows\u300d\u7684\u65e7\u6559\u7a0b\u3002\u7248\u672c\u8fed\u4ee3\u6781\u5feb&#xff0c;\u88c5\u4e4b\u524d\u5148 pip index versions &lt;pkg&gt; \u6216\u67e5 PyPI files \u9875\u786e\u8ba4\u6709\u6ca1\u6709 win_amd64 \u8f6e\u5b50&#xff0c;\u5426\u5219\u767d\u6298\u817e\u4e00\u4e0b\u5348\u3002<\/p>\n<hr \/>\n<h3 id=\"benchmark\">\u4e94\u3001\u6027\u80fd\u600e\u4e48\u6446&#xff08;\u516c\u5f00 benchmark&#xff0c;\u975e\u672c\u4eba\u5b9e\u6d4b&#xff09;<\/h3>\n<p>\u88c5\u5f97\u4e0a\u624d\u8c08\u6027\u80fd\u3002\u4e0b\u8868\u6c47\u603b\u5404\u9879\u76ee\u5b98\u65b9\u6587\u6863\/benchmark \u535a\u5ba2\u7684\u76f8\u5bf9\u7ed3\u8bba&#xff08;\u7edd\u5bf9 QPS \u968f\u786c\u4ef6\u5929\u5dee\u5730\u522b&#xff0c;\u8bf7\u5728\u4f60\u81ea\u5df1\u7684\u5361\u4e0a\u590d\u73b0&#xff09;&#xff1a;<\/p>\n<table>\n<tr>\u7ef4\u5ea6vLLMSGLangTensorRT-LLM<\/tr>\n<tbody>\n<tr>\n<td>\u541e\u5410(\u5bf9\u6bd4 HF)<\/td>\n<td>\u5b98\u65b9\u79f0\u8fde\u7eed\u6279\u5904\u7406\u5e26\u6765 2\u20134\u00d7 \u63d0\u5347<\/td>\n<td>\u524d\u7f00\u590d\u7528\u573a\u666f\u9ad8\u541e\u5410<\/td>\n<td>N \u5361\u4e0a\u901a\u5e38\u6700\u4f18<\/td>\n<\/tr>\n<tr>\n<td>TTFT(\u9996 token)<\/td>\n<td>\u4e2d<\/td>\n<td>\u4e2d<\/td>\n<td>\u4f4e&#xff08;\u7f16\u8bd1\u4f18\u5316&#xff09;<\/td>\n<\/tr>\n<tr>\n<td>TPOT(\u6bcf token)<\/td>\n<td>\u4f4e<\/td>\n<td>\u4f4e<\/td>\n<td>\u6700\u4f4e<\/td>\n<\/tr>\n<tr>\n<td>\u5cf0\u503c\u663e\u5b58<\/td>\n<td>\u4e2d&#xff08;PagedAttention&#xff09;<\/td>\n<td>\u4e2d<\/td>\n<td>\u4f4e&#xff08;\u5185\u6838\u878d\u5408&#xff09;<\/td>\n<\/tr>\n<tr>\n<td>\u6765\u6e90<\/td>\n<td>vLLM \u5b98\u65b9\u6587\u6863\/Blog<\/td>\n<td>SGLang GitHub<\/td>\n<td>TRT-LLM \u5b98\u65b9 Docs<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u6807\u6ce8&#xff1a;\u4e0a\u8868\u4e3a\u516c\u5f00\u6570\u636e&#xff08;\u975e\u672c\u4eba\u5b9e\u6d4b&#xff09;&#xff0c;\u6765\u6e90\u89c1\u6587\u672b\u3002\u7edd\u5bf9\u6570\u503c\u8bf7\u4ee5\u4f60\u672c\u673a\u538b\u6d4b\u4e3a\u51c6\u3002<\/p>\n<hr \/>\n<h3 id=\"_3\">\u516d\u3001\u8840\u6cea\u5751\u5b9e\u5f55<\/h3>\n<ul>\n<li>TRT-LLM \u9501 CUDA \u7248\u672c&#xff1a;\u7f16\u8bd1\u5bf9 CUDA\/driver \u6781\u5176\u6311\u5254&#xff0c;\u6362\u5f20\u5361\u6216\u5347\u9a71\u52a8\u53ef\u80fd\u6574\u5957\u91cd\u7f16&#xff0c;CI \u91cc\u662f\u5669\u68a6\u3002<\/li>\n<li>SGLang \u8def\u7531\u914d\u7f6e&#xff1a;RadixAttention \u5f3a&#xff0c;\u4f46\u591a\u5361\/\u8def\u7531\u53c2\u6570\u914d\u9519\u65f6\u541e\u5410\u4e0d\u5347\u53cd\u964d&#xff0c;\u65e5\u5fd7\u8fd8\u4e0d\u76f4\u89c2\u3002<\/li>\n<li>vLLM \u7684 prefix cache \u9ed8\u8ba4\u5173&#xff1a;\u591a\u8f6e\u5bf9\u8bdd\/\u5171\u4eab\u524d\u7f00\u573a\u666f&#xff0c;\u4e0d\u663e\u5f0f\u5f00 enable_prefix_caching \u7b49\u4e8e\u767d\u4e70\u5185\u5b58\u5e26\u5bbd\u3002<\/li>\n<\/ul>\n<hr \/>\n<h3 id=\"_4\">\u4e03\u3001\u9009\u578b\u51b3\u7b56\u6811<\/h3>\n<p>\u4f60\u7528\u7684\u662f N \u5361 &#043; Linux \u751f\u4ea7\u73af\u5883&#xff1f;<br \/>\n\u251c\u2500 \u662f&#xff0c;\u4e14\u8981\u6781\u81f4\u5ef6\u8fdf\/\u541e\u5410 \u2192 TensorRT-LLM&#xff08;\u80af\u82b1\u7f16\u8bd1\u6210\u672c&#xff09;<br \/>\n\u251c\u2500 \u662f&#xff0c;\u5e38\u89c4\u9ad8\u5e76\u53d1\u670d\u52a1     \u2192 vLLM&#xff08;\u6700\u7a33&#xff0c;\u793e\u533a\u7b54\u6848&#xff09;<br \/>\n\u2514\u2500 \u5426&#xff08;Windows \/ \u6d88\u8d39\u7ea7\u5361 \/ \u60f3\u5feb\u4e0a\u624b&#xff09;<br \/>\n   \u251c\u2500 \u591a\u8f6e\/Agent\/\u7ed3\u6784\u5316\u8f93\u51fa \u2192 \u4e0a WSL2 &#043; SGLang<br \/>\n   \u2514\u2500 \u901a\u7528\u63a8\u7406              \u2192 WSL2 &#043; vLLM<\/p>\n<p>\u4e00\u53e5\u8bdd&#xff1a;\u751f\u4ea7\u65e0\u8111 vLLM&#xff0c;N \u5361\u6781\u81f4\u4e0a TRT-LLM&#xff0c;\u7ed3\u6784\u5316\u8f93\u51fa\u770b SGLang&#xff1b;Windows \u7528\u6237\u5148\u88c5 WSL2&#xff0c;\u522b\u5728\u539f\u751f Windows \u91cc\u6b7b\u78d5\u3002<\/p>\n<hr \/>\n<h3 id=\"_5\">\u516b\u3001\u8bfb\u8005\u4ea4\u4ed8\u7269<\/h3>\n<p>\u2460 \u9009\u578b\u51b3\u7b56\u6811&#xff08;\u4e0a\u56fe&#xff0c;\u53ef\u622a\u56fe&#xff09; \u2461 \u4e00\u952e\u538b\u6d4b\u811a\u672c\u9aa8\u67b6&#xff08;vegeta \u7248&#xff0c;GitHub \u98ce\u683c\u7247\u6bb5&#xff09;&#xff1a;<\/p>\n<p># \u7528 vegeta \u538b\u4f60\u7684 \/v1\/completions \u63a5\u53e3<br \/>\necho &#039;{&#034;model&#034;:&#034;qwen2.5&#034;,&#034;prompt&#034;:&#034;\u4f60\u597d&#034;,&#034;max_tokens&#034;:128}&#039; \\\\<br \/>\n  | vegeta attack -rate&#061;20 -duration&#061;30s -method&#061;POST \\\\<br \/>\n    -header&#061;&#034;Content-Type: application\/json&#034; \\\\<br \/>\n    -target&#061;https:\/\/your-host\/v1\/completions \\\\<br \/>\n  | vegeta report      # \u770b\u541e\u5410 QPS \/ p99 \u5ef6\u8fdf<\/p>\n<hr \/>\n<h3 id=\"3\">\u7ed3\u5c3e&#xff1a;\u4eca\u5929\u5c31\u80fd\u52a8\u624b\u7684 3 \u6761\u6e05\u5355<\/h3>\n<li>\u5148\u67e5\u8f6e\u5b50\u518d\u52a8\u624b&#xff1a;pip index versions vllm&#xff0c;\u786e\u8ba4\u6709\u4f60\u7cfb\u7edf\u7684 wheel&#xff0c;\u522b\u4e0a\u6765\u5c31 pip install\u3002<\/li>\n<li>\u88c5 WSL2&#xff1a;Windows \u7528\u6237\u82b1 10 \u5206\u949f wsl &#8211;install&#xff0c;\u63a8\u7406\u5f15\u64ce\u7684\u4e16\u754c\u5728 Linux \u4fa7\u3002<\/li>\n<li>\u538b\u4e00\u6b21\u57fa\u7ebf&#xff1a;\u7528\u4e0a\u9762 vegeta \u7247\u6bb5\u538b\u4f60\u73b0\u6709\u63a5\u53e3&#xff0c;\u8bb0\u4e0b QPS\/p99&#xff0c;\u4e0b\u7bc7\u6211\u4eec\u62c6\u5ef6\u8fdf\u3002<\/li>\n<hr \/>\n<p>\u5b9e\u6d4b\u73af\u5883\u4e0e\u514d\u8d23&#xff1a;\u672c\u6b21\u73af\u5883\u89c1\u7b2c\u4e09\u8282\u3002\u5b89\u88c5\u53ef\u88c5\u6027\u4e3a\u672c\u4eba\u5728\u672c\u673a\/PyPI \u5b9e\u8dd1&#xff1b;\u6027\u80fd\u541e\u5410\u4e3a\u5404\u9879\u76ee\u5b98\u65b9\u516c\u5f00 benchmark&#xff08;\u975e\u672c\u4eba\u5b9e\u6d4b&#xff09;&#xff0c;\u6765\u6e90&#xff1a;vLLM \u5b98\u65b9\u6587\u6863\u4e0e Blog\u3001SGLang GitHub README\u3001NVIDIA TensorRT-LLM \u5b98\u65b9 Docs\u3002\u6570\u636e\u57fa\u4e8e\u672c\u4eba\u73af\u5883&#xff0c;\u4ec5\u4f9b\u53c2\u8003&#xff0c;\u8bf7\u4ee5\u4f60\u81ea\u5df1\u7684\u786c\u4ef6\u590d\u6d4b\u4e3a\u51c6\u3002<\/p>\n<p>\u4e0b\u7bc7\u6211\u4eec\u804a&#xff1a;\u4e3a\u4ec0\u4e48\u4e09\u5f15\u64ce\u541e\u5410\u5dee\u8fd9\u4e48\u591a&#xff1f;\u2014\u2014\u8fde\u7eed\u6279\u5904\u7406 &#043; \u6295\u673a\u89e3\u7801\u7684\u5b9e\u6d4b\u8d26\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u7cfb\u5217\u7b2c 1 \u7bc7 \u00b7 \u63a8\u7406\u9009\u578b\u51b3\u7b56 \u5b98\u7f51\u8bf4 vLLM \u6bd4 HuggingFace Transformers \u5feb 3 \u500d&#xff0c;\u4f46\u6211\u5728\u81ea\u5df1\u7684 RTX 5060 \u7b14\u8bb0\u672c\u4e0a\u6572 pip install vllm&#xff0c;\u76f4\u63a5\u62a5\u300c\u627e\u4e0d\u5230\u5339\u914d\u7684 Windows \u5305\u300d\u3002\u5148\u522b\u76f2\u4fe1 benchmark&#xff0c;\u90e8\u7f72\u95e8\u69db\u624d\u662f\u7b2c\u4e00\u9053\u5173\u3002<br \/>\n\u672c\u8f91\u4e3b\u7ebf&#xff1a;\u628a\u5927\u6a21\u578b\u4ece\u300c\u80fd\u8c03\u901a\u300d\u505a\u5230\u300c\u8dd1\u5f97\u5feb\u3001\u8dd1\u5f97\u7701\u3001\u6d4b\u5f97\u51c6\u3001\u5b88\u5f97\u4f4f\u300d\u3002\u8fd9\u4e00\u7bc7\u5148\u89e3\u51b3\u9009\u578b\u2014\u2014\u540c\u4e00\u6a21\u578b\u3001\u540c\u4e00\u786c\u4ef6&#xff0c;\u4e09\u4e2a<\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[10162,6994,2681,149,132],"topic":[],"class_list":["post-89586","post","type-post","status-publish","format-standard","hentry","category-server","tag-ai","tag-sglang","tag-vllm","tag-windows","tag-132"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v20.3 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>vLLM \/ TensorRT-LLM \/ SGLang \u6a2a\u8bc4\uff1a\u6211\u5728 Windows \u4e0a\u8fde pip install \u90fd\u88c5\u4e0d\u4e0a - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.wsisp.com\/helps\/89586.html\" \/>\n<meta property=\"og:locale\" content=\"zh_CN\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"vLLM \/ TensorRT-LLM \/ SGLang \u6a2a\u8bc4\uff1a\u6211\u5728 Windows \u4e0a\u8fde pip install \u90fd\u88c5\u4e0d\u4e0a - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"og:description\" content=\"\u7cfb\u5217\u7b2c 1 \u7bc7 \u00b7 \u63a8\u7406\u9009\u578b\u51b3\u7b56 \u5b98\u7f51\u8bf4 vLLM \u6bd4 HuggingFace Transformers \u5feb 3 \u500d&#xff0c;\u4f46\u6211\u5728\u81ea\u5df1\u7684 RTX 5060 \u7b14\u8bb0\u672c\u4e0a\u6572 pip install vllm&#xff0c;\u76f4\u63a5\u62a5\u300c\u627e\u4e0d\u5230\u5339\u914d\u7684 Windows \u5305\u300d\u3002\u5148\u522b\u76f2\u4fe1 benchmark&#xff0c;\u90e8\u7f72\u95e8\u69db\u624d\u662f\u7b2c\u4e00\u9053\u5173\u3002 \u672c\u8f91\u4e3b\u7ebf&#xff1a;\u628a\u5927\u6a21\u578b\u4ece\u300c\u80fd\u8c03\u901a\u300d\u505a\u5230\u300c\u8dd1\u5f97\u5feb\u3001\u8dd1\u5f97\u7701\u3001\u6d4b\u5f97\u51c6\u3001\u5b88\u5f97\u4f4f\u300d\u3002\u8fd9\u4e00\u7bc7\u5148\u89e3\u51b3\u9009\u578b\u2014\u2014\u540c\u4e00\u6a21\u578b\u3001\u540c\u4e00\u786c\u4ef6&#xff0c;\u4e09\u4e2a\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.wsisp.com\/helps\/89586.html\" \/>\n<meta property=\"og:site_name\" content=\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"article:published_time\" content=\"2026-08-03T01:57:25+00:00\" \/>\n<meta name=\"author\" content=\"admin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u4f5c\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin\" \/>\n\t<meta name=\"twitter:label2\" content=\"\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4\" \/>\n\t<meta name=\"twitter:data2\" content=\"2 \u5206\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/89586.html\",\"url\":\"https:\/\/www.wsisp.com\/helps\/89586.html\",\"name\":\"vLLM \/ TensorRT-LLM \/ SGLang \u6a2a\u8bc4\uff1a\u6211\u5728 Windows \u4e0a\u8fde pip install \u90fd\u88c5\u4e0d\u4e0a - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"isPartOf\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\"},\"datePublished\":\"2026-08-03T01:57:25+00:00\",\"dateModified\":\"2026-08-03T01:57:25+00:00\",\"author\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\"},\"breadcrumb\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/89586.html#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.wsisp.com\/helps\/89586.html\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/89586.html#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\/\/www.wsisp.com\/helps\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"vLLM \/ TensorRT-LLM \/ SGLang \u6a2a\u8bc4\uff1a\u6211\u5728 Windows \u4e0a\u8fde pip install \u90fd\u88c5\u4e0d\u4e0a\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\",\"url\":\"https:\/\/www.wsisp.com\/helps\/\",\"name\":\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"description\":\"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\",\"name\":\"admin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"contentUrl\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"caption\":\"admin\"},\"sameAs\":[\"http:\/\/wp.wsisp.com\"],\"url\":\"https:\/\/www.wsisp.com\/helps\/author\/admin\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"vLLM \/ TensorRT-LLM \/ SGLang \u6a2a\u8bc4\uff1a\u6211\u5728 Windows \u4e0a\u8fde pip install \u90fd\u88c5\u4e0d\u4e0a - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.wsisp.com\/helps\/89586.html","og_locale":"zh_CN","og_type":"article","og_title":"vLLM \/ TensorRT-LLM \/ SGLang \u6a2a\u8bc4\uff1a\u6211\u5728 Windows \u4e0a\u8fde pip install \u90fd\u88c5\u4e0d\u4e0a - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","og_description":"\u7cfb\u5217\u7b2c 1 \u7bc7 \u00b7 \u63a8\u7406\u9009\u578b\u51b3\u7b56 \u5b98\u7f51\u8bf4 vLLM \u6bd4 HuggingFace Transformers \u5feb 3 \u500d&#xff0c;\u4f46\u6211\u5728\u81ea\u5df1\u7684 RTX 5060 \u7b14\u8bb0\u672c\u4e0a\u6572 pip install vllm&#xff0c;\u76f4\u63a5\u62a5\u300c\u627e\u4e0d\u5230\u5339\u914d\u7684 Windows \u5305\u300d\u3002\u5148\u522b\u76f2\u4fe1 benchmark&#xff0c;\u90e8\u7f72\u95e8\u69db\u624d\u662f\u7b2c\u4e00\u9053\u5173\u3002 \u672c\u8f91\u4e3b\u7ebf&#xff1a;\u628a\u5927\u6a21\u578b\u4ece\u300c\u80fd\u8c03\u901a\u300d\u505a\u5230\u300c\u8dd1\u5f97\u5feb\u3001\u8dd1\u5f97\u7701\u3001\u6d4b\u5f97\u51c6\u3001\u5b88\u5f97\u4f4f\u300d\u3002\u8fd9\u4e00\u7bc7\u5148\u89e3\u51b3\u9009\u578b\u2014\u2014\u540c\u4e00\u6a21\u578b\u3001\u540c\u4e00\u786c\u4ef6&#xff0c;\u4e09\u4e2a","og_url":"https:\/\/www.wsisp.com\/helps\/89586.html","og_site_name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","article_published_time":"2026-08-03T01:57:25+00:00","author":"admin","twitter_card":"summary_large_image","twitter_misc":{"\u4f5c\u8005":"admin","\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4":"2 \u5206"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.wsisp.com\/helps\/89586.html","url":"https:\/\/www.wsisp.com\/helps\/89586.html","name":"vLLM \/ TensorRT-LLM \/ SGLang \u6a2a\u8bc4\uff1a\u6211\u5728 Windows \u4e0a\u8fde pip install \u90fd\u88c5\u4e0d\u4e0a - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","isPartOf":{"@id":"https:\/\/www.wsisp.com\/helps\/#website"},"datePublished":"2026-08-03T01:57:25+00:00","dateModified":"2026-08-03T01:57:25+00:00","author":{"@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41"},"breadcrumb":{"@id":"https:\/\/www.wsisp.com\/helps\/89586.html#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.wsisp.com\/helps\/89586.html"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/www.wsisp.com\/helps\/89586.html#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/www.wsisp.com\/helps"},{"@type":"ListItem","position":2,"name":"vLLM \/ TensorRT-LLM \/ SGLang \u6a2a\u8bc4\uff1a\u6211\u5728 Windows \u4e0a\u8fde pip install \u90fd\u88c5\u4e0d\u4e0a"}]},{"@type":"WebSite","@id":"https:\/\/www.wsisp.com\/helps\/#website","url":"https:\/\/www.wsisp.com\/helps\/","name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","description":"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"zh-Hans"},{"@type":"Person","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41","name":"admin","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/","url":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","contentUrl":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","caption":"admin"},"sameAs":["http:\/\/wp.wsisp.com"],"url":"https:\/\/www.wsisp.com\/helps\/author\/admin"}]}},"_links":{"self":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/89586","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/comments?post=89586"}],"version-history":[{"count":0,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/89586\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media?parent=89586"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/categories?post=89586"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/tags?post=89586"},{"taxonomy":"topic","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/topic?post=89586"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}