{"id":206,"date":"2024-07-25T21:16:43","date_gmt":"2024-07-25T13:16:43","guid":{"rendered":"https:\/\/www.onetts.com\/ai\/?p=206"},"modified":"2024-08-01T14:35:15","modified_gmt":"2024-08-01T06:35:15","slug":"deepseek-v2-chat","status":"publish","type":"post","link":"https:\/\/www.onetts.com\/ai\/models\/deepseek-v2-chat\/","title":{"rendered":"DeepSeek-V2-Chat"},"content":{"rendered":"<p>DeepSeek-V2-Chat \u662f DeepSeek AI \u63a8\u51fa\u7684\u4e00\u6b3e\u57fa\u4e8e\u4e13\u5bb6\u6df7\u5408\uff08Mixture-of-Experts, MoE\uff09\u7684\u5927\u8bed\u8a00\u6a21\u578b\uff0c\u5177\u6709\u5f3a\u5927\u7684\u8bed\u8a00\u7406\u89e3\u4e0e\u751f\u6210\u80fd\u529b\u3002\u8be5\u6a21\u578b\u62e5\u6709 2360 \u4ebf\u4e2a\u53c2\u6570\uff0c\u5176\u4e2d\u6bcf\u4e2a token \u6fc0\u6d3b\u4e86 210 \u4ebf\u4e2a\u53c2\u6570\u3002DeepSeek-V2 \u5728\u4fdd\u6301\u6027\u80fd\u63d0\u5347\u7684\u540c\u65f6\uff0c\u5927\u5e45\u5ea6\u964d\u4f4e\u4e86\u8bad\u7ec3\u6210\u672c\uff0c\u51cf\u5c11\u4e86\u952e\u503c\u7f13\u5b58\u7684\u4f7f\u7528\uff0c\u5e76\u663e\u8457\u63d0\u9ad8\u4e86\u751f\u6210\u541e\u5410\u91cf\u3002<\/p>\n<h2>\u6a21\u578b\u8bc4\u6d4b<\/h2>\n<p>DeepSeek-V2-Chat \u5728\u591a\u4e2a\u6807\u51c6\u57fa\u51c6\u6d4b\u8bd5\u4e2d\u5c55\u73b0\u51fa\u5353\u8d8a\u7684\u6027\u80fd\uff0c\u5305\u62ec\u4f46\u4e0d\u9650\u4e8e MMLU\u3001BBH\u3001C-Eval\u3001CMMLU\u3001HumanEval\u3001MBPP\u3001GSM8K \u548c Math \u7b49\u3002\u5728\u82f1\u8bed\u548c\u4e2d\u6587\u7684\u5bf9\u8bdd\u751f\u6210\u8bc4\u4f30\u4e2d\uff0cDeepSeek-V2-Chat \u540c\u6837\u8868\u73b0\u51fa\u8272\uff0c\u5176\u5728 AlpacaEval 2.0\u3001MTBench \u548c Alignbench \u4e0a\u7684\u6210\u7ee9\u663e\u793a\u4e86\u5176\u5728\u5bf9\u8bdd\u751f\u6210\u65b9\u9762\u7684\u7ade\u4e89\u529b\u3002<\/p>\n<h2>\u90e8\u7f72\u4f7f\u7528<\/h2>\n<h3>\u90e8\u7f72\u6b65\u9aa4<\/h3>\n<ol>\n<li><strong>\u73af\u5883\u51c6\u5907<\/strong>\uff1a\u786e\u4fdd\u7cfb\u7edf\u5b89\u88c5\u4e86 Python \u73af\u5883\u4ee5\u53ca\u5fc5\u8981\u7684\u5e93\uff0c\u5982 PyTorch \u548c Huggingface \u7684 Transformers\u3002<\/li>\n<li><strong>\u6a21\u578b\u4e0b\u8f7d<\/strong>\uff1a\u4ece DeepSeek AI \u7684\u5b98\u65b9\u8d44\u6e90\u4e0b\u8f7d DeepSeek-V2-Chat \u6a21\u578b\u3002<\/li>\n<li><strong>\u5b89\u88c5\u4f9d\u8d56<\/strong>\uff1a\u4f7f\u7528 pip \u5b89\u88c5\u6240\u9700\u7684 Python \u5305\u3002<\/li>\n<li><strong>\u6a21\u578b\u52a0\u8f7d<\/strong>\uff1a\u4f7f\u7528 Huggingface \u7684 Transformers \u52a0\u8f7d\u6a21\u578b\u548c\u5206\u8bcd\u5668\u3002<\/li>\n<li><strong>\u6a21\u578b\u63a8\u7406<\/strong>\uff1a\u6839\u636e\u4e1a\u52a1\u9700\u6c42\u7f16\u5199\u63a8\u7406\u4ee3\u7801\uff0c\u8c03\u7528\u6a21\u578b\u8fdb\u884c\u6587\u672c\u751f\u6210\u6216\u5bf9\u8bdd\u4ea4\u4e92\u3002<\/li>\n<\/ol>\n<h2>\u5e38\u89c1\u95ee\u9898\u4e0e\u89e3\u7b54<\/h2>\n<p><strong>Q: \u5982\u4f55\u4f18\u5316\u6a21\u578b\u5728 GPU \u4e0a\u7684\u6027\u80fd\uff1f<\/strong><\/p>\n<p>A: \u63a8\u8350\u4f7f\u7528 BF16 \u683c\u5f0f\u8fdb\u884c\u63a8\u7406\uff0c\u9700\u8981 80GB*8 \u7684 GPU \u8d44\u6e90\u3002\u53ef\u4ee5\u901a\u8fc7\u8c03\u6574 <code class=\"inlineCode___uUrJ9\">max_memory<\/code> \u53c2\u6570\u6765\u4f18\u5316\u5185\u5b58\u4f7f\u7528\u3002<\/p>\n<p><strong>Q: \u5982\u4f55\u5904\u7406\u6a21\u578b\u751f\u6210\u7684\u6587\u672c\uff1f<\/strong><\/p>\n<p>A: \u4f7f\u7528\u5206\u8bcd\u5668\u7684 <code class=\"inlineCode___uUrJ9\">decode<\/code> \u65b9\u6cd5\u5c06\u751f\u6210\u7684 token \u5e8f\u5217\u8f6c\u6362\u56de\u53ef\u8bfb\u6587\u672c\u3002<\/p>\n<p><strong>Q: \u5982\u4f55\u5728\u672c\u5730\u8fd0\u884c DeepSeek-V2-Chat \u8fdb\u884c\u5bf9\u8bdd\uff1f<\/strong><\/p>\n<p>A: \u53ef\u4ee5\u53c2\u8003 DeepSeek AI \u63d0\u4f9b\u7684\u804a\u5929\u5b8c\u6210\u793a\u4f8b\u4ee3\u7801\uff0c\u8be5\u4ee3\u7801\u5c55\u793a\u4e86\u5982\u4f55\u4f7f\u7528\u6a21\u578b\u8fdb\u884c\u5bf9\u8bdd\u751f\u6210\u3002<\/p>\n<h2>\u76f8\u5173\u8d44\u6e90<\/h2>\n<ul>\n<li><strong>\u5b98\u65b9\u6a21\u578b\u9875\u9762<\/strong>\uff1a<a target=\"_blank\" rel=\"noopener noreferrer nofollow\" href=\"https:\/\/www.modelscope.cn\/models\/deepseek-ai\/DeepSeek-V2-Chat\" target=\"_blank\" rel=\"noopener\">DeepSeek-V2-Chat on ModelScope<\/a><\/li>\n<li><strong>GitHub \u4ed3\u5e93<\/strong>\uff1a<a target=\"_blank\" rel=\"noopener noreferrer nofollow\" href=\"https:\/\/github.com\/deepseek-ai\/DeepSeek-V2\" target=\"_blank\" rel=\"noopener\">DeepSeek-V2 GitHub<\/a><\/li>\n<li><strong>\u5b98\u65b9\u804a\u5929\u7f51\u7ad9<\/strong>\uff1a<a target=\"_blank\" rel=\"noopener noreferrer nofollow\" href=\"https:\/\/chat.deepseek.com\/\" target=\"_blank\" rel=\"noopener\">chat.deepseek.com<\/a><\/li>\n<li><strong>API \u5e73\u53f0<\/strong>\uff1a<a target=\"_blank\" rel=\"noopener noreferrer nofollow\" href=\"https:\/\/platform.deepseek.com\/\" target=\"_blank\" rel=\"noopener\">platform.deepseek.com<\/a><\/li>\n<li><strong>\u6280\u672f\u6587\u6863<\/strong>\uff1a\u53ef\u4ee5\u5728 DeepSeek AI \u7684\u5b98\u65b9\u6587\u6863\u4e2d\u627e\u5230\u66f4\u591a\u5173\u4e8e\u6a21\u578b\u90e8\u7f72\u548c\u4f7f\u7528\u7684\u8be6\u7ec6\u4fe1\u606f\u3002<\/li>\n<\/ul>\n<h3>\u7ed3\u8bed<\/h3>\n<p class=\"last-node\">DeepSeek-V2-Chat \u4ee5\u5176\u9ad8\u6548\u7684\u6027\u80fd\u548c\u5e7f\u6cdb\u7684\u5e94\u7528\u524d\u666f\uff0c\u5728\u81ea\u7136\u8bed\u8a00\u5904\u7406\u9886\u57df\u5c55\u73b0\u51fa\u5de8\u5927\u7684\u6f5c\u529b\u3002\u968f\u7740\u6280\u672f\u7684\u4e0d\u65ad\u8fdb\u6b65\u548c\u793e\u533a\u7684\u5171\u540c\u52aa\u529b\uff0c\u671f\u5f85 DeepSeek-V2-Chat \u80fd\u591f\u5728\u667a\u80fd\u5bf9\u8bdd\u3001\u5185\u5bb9\u751f\u6210\u7b49\u591a\u4e2a\u9886\u57df\u53d1\u6325\u66f4\u5927\u7684\u4f5c\u7528\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"<p>DeepSeek-V2-Chat \u662f DeepSeek AI \u63a8\u51fa\u7684\u4e00\u6b3e\u57fa\u4e8e\u4e13\u5bb6\u6df7\u5408\uff08Mixture-of-Experts, MoE\uff09\u7684\u5927\u8bed\u8a00\u6a21\u578b\uff0c\u5177\u6709\u5f3a\u5927\u7684\u8bed\u8a00\u7406\u89e3\u4e0e\u751f\u6210\u80fd\u529b\u3002\u8be5\u6a21\u578b\u62e5\u6709 2360 \u4ebf\u4e2a\u53c2\u6570\uff0c\u5176\u4e2d\u6bcf\u4e2a token \u6fc0\u6d3b\u4e86 210 \u4ebf\u4e2a\u53c2\u6570\u3002DeepSeek-V2 \u5728\u4fdd\u6301\u6027\u80fd\u63d0\u5347\u7684\u540c\u65f6\uff0c\u5927\u5e45\u5ea6\u964d\u4f4e\u4e86\u8bad\u7ec3\u6210\u672c\uff0c\u51cf\u5c11\u4e86\u952e\u503c\u7f13\u5b58\u7684\u4f7f\u7528\uff0c\u5e76\u663e\u8457\u63d0\u9ad8\u4e86\u751f\u6210\u541e\u5410\u91cf\u3002 \u6a21\u578b\u8bc4\u6d4b DeepSeek-V2-Chat \u5728\u591a\u4e2a\u6807\u51c6\u57fa\u51c6\u6d4b\u8bd5\u4e2d\u5c55\u73b0\u51fa\u5353\u8d8a\u7684\u6027\u80fd\uff0c\u5305\u62ec\u4f46\u4e0d\u9650\u4e8e MMLU\u3001BBH\u3001C-Eval\u3001CMMLU\u3001Hum<\/p>\n","protected":false},"author":1,"featured_media":207,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3],"tags":[43],"collection":[],"company":[24],"rank":[54,53],"class_list":["post-206","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-llm","tag-moe-llm","company-deepseek","rank-llm-leaderboard","rank-china-llm-ranking"],"_links":{"self":[{"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/posts\/206","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/comments?post=206"}],"version-history":[{"count":1,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/posts\/206\/revisions"}],"predecessor-version":[{"id":208,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/posts\/206\/revisions\/208"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/media\/207"}],"wp:attachment":[{"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/media?parent=206"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/categories?post=206"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/tags?post=206"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/collection?post=206"},{"taxonomy":"company","embeddable":true,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/company?post=206"},{"taxonomy":"rank","embeddable":true,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/rank?post=206"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}