{"id":269,"date":"2025-12-07T14:27:27","date_gmt":"2025-12-07T06:27:27","guid":{"rendered":"https:\/\/blog.zeronely.cn\/?p=269"},"modified":"2025-12-07T14:27:28","modified_gmt":"2025-12-07T06:27:28","slug":"%e6%9f%90%e7%a7%8d%e6%8f%90%e5%8f%96%e5%85%b3%e8%81%94%e7%9a%84%e6%96%b9%e6%a1%88%e6%b1%87%e6%8a%a5","status":"publish","type":"post","link":"https:\/\/blog.zeronely.cn\/index.php\/2025\/12\/07\/%e6%9f%90%e7%a7%8d%e6%8f%90%e5%8f%96%e5%85%b3%e8%81%94%e7%9a%84%e6%96%b9%e6%a1%88%e6%b1%87%e6%8a%a5\/","title":{"rendered":"\u67d0\u79cd\u63d0\u53d6\u5173\u8054\u7684\u65b9\u6848\u6c47\u62a5"},"content":{"rendered":"\n

\u73af\u58833090\uff0824g\uff09*2<\/p>\n\n\n\n

\u6a21\u578bqwen1.5-14b-chat<\/p>\n\n\n\n

\u53c2\u6570<\/p>\n\n\n\n

python -m vllm.entrypoints.openai.api_server<\/p>\n\n\n\n

–model \/root\/models\/qwen1.5-14b-chat<\/p>\n\n\n\n

–tensor-parallel-size 2 #\u53cc\u5361\u5e76\u884c<\/p>\n\n\n\n

–trust-remote-code<\/p>\n\n\n\n

–host 0.0.0.0<\/p>\n\n\n\n

–port 8000<\/p>\n\n\n\n

–gpu-memory-utilization 0.9 #\u5b58\u7559\u663e\u5b58\u4f59\u91cf<\/p>\n\n\n\n

–dtype bfloat16<\/p>\n\n\n\n

–max-model-len 4096\uff08Qwen1.5-14B #\u9ed8\u8ba4\u652f\u6301 32768 \u4e2a Token \u7684\u8d85\u957f\u4e0a\u4e0b\u6587\uff08\u7ea6 24000 \u5b57\uff09\uff0c\u628a max_model_len \u964d\u5230 4096\uff08\u7ea6 3000 \u5b57\uff09\uff0c\u5185\u5b58\u4e0d\u8db3\uff0c\u6a21\u578b\u5206\u914d\u5b8c\u540e\u53ea\u52695.27g\uff09<\/p>\n\n\n\n

\u51e0\u79cd\u5c1d\u8bd5\uff1a<\/p>\n\n\n\n

\u8f93\u5165\u5185\u5bb9\u6539\u53d8\uff0c\u7531\u7b80\u5316\u8f68\u8ff9\u53d8\u6210\u5b8c\u6574\u7684\u8f68\u8ff9\u70b9\u5e8f\u5217\uff1b<\/p>\n\n\n\n

\u660e\u786e\u8f93\u51fa\u683c\u5f0f\u8981\u6c42\uff0c\u9650\u5b9a\u5173\u7cfb\u7c7b\u578b\u4e3a \u201c\u8ddf\u968f \/ \u907f\u8ba9 \/ \u65e0\u660e\u663e\u5173\u7cfb\u201d\uff1b<\/p>\n\n\n\n

\u8865\u5145\u91cf\u5316\u5173\u7cfb\u5b9a\u4e49\uff08\u5982\u8ddf\u968f\u5173\u7cfb\uff1a\u8ddd\u79bb 1-2 \u6d77\u91cc + \u822a\u5411\u504f\u5dee\u226410\u00b0+ \u901f\u5ea6\u5dee\u22642 \u8282\uff09\uff1b<\/p>\n\n\n\n

\u4f18\u5316\u8f68\u8ff9\u70b9\u63cf\u8ff0\u683c\u5f0f\uff08\u53bb\u6389\u5197\u4f59\u5206\u9694\u7b26\uff0c\u6807\u51c6\u5316\u5b57\u6bb5\u987a\u5e8f\uff09\uff0c\u5206\u5757\u8f93\u5165\u8bbe\u8ba1\uff1b<\/p>\n\n\n\n

\u964d\u4f4e\u6e29\u5ea6\u53c2\u6570\uff08temperature=0.2\uff09\uff0c\u589e\u52a0\u9519\u8bef\u793a\u4f8b\u5bf9\u6bd4\uff0c\u5f3a\u5316\u7ea6\u675f\u8bed\u6c14\uff1b<\/p>\n\n\n\n

\u4fee\u6539\u63d0\u793a\u8bcd\uff0c\u5c1d\u8bd5\u964d\u4f4e\u7ea6\u675f\uff1b<\/p>\n\n\n\n

\u7ed3\u679c\uff1a\u76ee\u524d\u5728**\u5c0f\u89c4\u6a21\u7684\u6a21\u62df\u6570\u636e**<\/strong>\u4e2d\u6548\u679c\u8f83\u597d\u3002<\/p>\n\n\n\n

\u7531\u4e8e\u6a21\u578b\u4e0a\u4e0b\u6587\u957f\u5ea6\u6709\u9650\uff0c\u91c7\u53d6\u63aa\u65bd\u91cd\u65b0\u6e05\u6d17\u6570\u636e<\/p>\n\n\n\n

\u7edf\u4e00\u5b57\u6bb5\u540d\u3001\u6807\u51c6\u5316\u65f6\u95f4\u683c\u5f0f<\/p>\n\n\n\n

\u5220\u9664\u7f3a\u5931\u5173\u952e\u4fe1\u606f\uff08MMSI\u3001\u7ecf\u7eac\u5ea6\u3001\u901f\u5ea6\u7b49\uff09\u3001\u7269\u7406\u5f02\u5e38\uff08\u8d85\u5e38\u89c4\u901f\u5ea6 \/ \u822a\u5411 \/ \u7ecf\u7eac\u5ea6\uff09\u3001\u91cd\u590d\u8bb0\u5f55\u7684\u6570\u636e<\/p>\n\n\n\n

\u7cbe\u7b80\u683c\u5f0f\uff08\u4fdd\u7559\u5408\u7406\u5c0f\u6570\u4f4d\uff09\uff0c\u51cf\u5c11\u5197\u4f59\u6570\u636e\u5360\u7528\u7684 Token \u91cf<\/p>\n\n\n\n

\u91c7\u6837\u65f6\u91c7\u7528**\u5173\u952e\u8282\u70b9\u548c\u5747\u5300\u91c7\u6837**<\/strong>\u62fc\u63a5\uff0c\u907f\u514d\u8fc7\u4e8e\u5bc6\u96c6\uff0c\u540c\u65f6\u4fdd\u7559\u884c\u4e3a\u7279\u5f81<\/p>\n\n\n\n

\u6b64\u5916\uff0c\u8fdb\u884c\u6309\u7167\u65f6\u7a7a\u5207\u5206\uff0c\u786e\u4fdd\u540c\u4e00\u7a97\u53e3\u5185\u7684\u8239\u8236\u96c6\u7fa4\u65f6\u95f4\u540c\u6b65\u3001\u7a7a\u95f4\u8ddd\u79bb\u5408\u9002<\/p>\n\n\n\n

\u6700\u7ec8\u751f\u6210\u6700\u5c0f\u5206\u6790\u5355\u5143–\u8239\u5bf9<\/p>\n\n\n\n

\u5c06\u8239\u961f\u7528\u6279\u5904\u7406\u811a\u672c\u8f93\u5165\u5927\u6a21\u578b\uff0c\u5224\u65ad\u4e24\u4e24\u4e4b\u95f4\u7684\u5173\u7cfb<\/p>\n\n\n\n

\u5927\u591a\u6570\u65e0\u660e\u663e\u5173\u7cfb<\/p>\n","protected":false},"excerpt":{"rendered":"

\u73af\u58833090\uff0824g\uff09*2 \u6a21\u578bqwen1.5-14b-chat \u53c2\u6570 python -m vllm.entr […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-269","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/posts\/269","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/comments?post=269"}],"version-history":[{"count":1,"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/posts\/269\/revisions"}],"predecessor-version":[{"id":270,"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/posts\/269\/revisions\/270"}],"wp:attachment":[{"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/media?parent=269"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/categories?post=269"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/tags?post=269"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}