{"id":209,"date":"2024-07-25T21:25:39","date_gmt":"2024-07-25T13:25:39","guid":{"rendered":"https:\/\/www.onetts.com\/ai\/?p=209"},"modified":"2024-08-01T14:34:19","modified_gmt":"2024-08-01T06:34:19","slug":"qwen1-5-110b-chat","status":"publish","type":"post","link":"https:\/\/www.onetts.com\/ai\/models\/qwen1-5-110b-chat\/","title":{"rendered":"Qwen1.5-110B-Chat"},"content":{"rendered":"<p data-spm-anchor-id=\"a2c6h.13066512.0.i3.5bcd36afk11UPI\">Qwen1.5-110B-Chat \u662f Qwen2 \u7cfb\u5217\u7684\u6d4b\u8bd5\u7248\uff0c\u5c5e\u4e8e\u57fa\u4e8e Transformer \u67b6\u6784\u7684\u89e3\u7801\u5668\u4ec5\u8bed\u8a00\u6a21\u578b\u3002\u5b83\u7531\u4e0d\u540c\u5c3a\u5bf8\u7684\u6a21\u578b\u7ec4\u6210\uff0c\u4ece 0.5B \u5230 110B\uff0c\u5305\u62ec\u4e00\u4e2a 14B \u7684 MoE \u6a21\u578b\uff0c\u5177\u6709 2.7B \u7684\u6fc0\u6d3b\u91cf\u3002\u8be5\u6a21\u578b\u5728\u4eba\u7c7b\u504f\u597d\u3001\u591a\u8bed\u8a00\u652f\u6301\u3001\u4e0a\u4e0b\u6587\u957f\u5ea6\u7a33\u5b9a\u6027\u7b49\u65b9\u9762\u8fdb\u884c\u4e86\u663e\u8457\u6539\u8fdb\uff0c\u5e76\u652f\u6301\u9ad8\u8fbe 32K \u7684\u4e0a\u4e0b\u6587\u957f\u5ea6\u3002<\/p>\n<h2 id=\"-\">\u6a21\u578b\u8bc4\u6d4b<\/h2>\n<p>Qwen1.5-110B-Chat \u5728\u591a\u4e2a\u7ef4\u5ea6\u4e0a\u8fdb\u884c\u4e86\u6027\u80fd\u8bc4\u6d4b\uff0c\u5305\u62ec\u4f46\u4e0d\u9650\u4e8e\u8bed\u8a00\u751f\u6210\u80fd\u529b\u3001\u591a\u8bed\u8a00\u7406\u89e3\u548c\u751f\u6210\u3001\u4ee5\u53ca\u5728\u7279\u5b9a\u4efb\u52a1\u4e0a\u7684\u8868\u73b0\u3002\u6a21\u578b\u7684\u8bc4\u6d4b\u7ed3\u679c\u8868\u660e\uff0c\u5b83\u5728\u804a\u5929\u6a21\u578b\u7684\u4eba\u7c7b\u504f\u597d\u65b9\u9762\u6709\u663e\u8457\u63d0\u5347\uff0c\u5e76\u4e14\u5728\u591a\u8bed\u8a00\u652f\u6301\u65b9\u9762\u8868\u73b0\u51fa\u8272\u3002<\/p>\n<h2 id=\"-\">\u90e8\u7f72\u4f7f\u7528<\/h2>\n<h3 id=\"-\">\u90e8\u7f72\u6b65\u9aa4<\/h3>\n<ol>\n<li><strong>\u5b89\u88c5\u4f9d\u8d56<\/strong>\uff1a\u786e\u4fdd\u5df2\u5b89\u88c5 Python \u548c PyTorch\u3002\u63a8\u8350\u5b89\u88c5 Hugging Face \u7684\u00a0<code>transformers<\/code>\u00a0\u5e93\uff0c\u7248\u672c\u9700\u5927\u4e8e\u7b49\u4e8e 4.37.0\u3002<\/li>\n<li><strong>\u52a0\u8f7d\u6a21\u578b\u548c\u5206\u8bcd\u5668<\/strong>\uff1a\n<pre><code class=\"lang-python\"><span class=\"token keyword\">from<\/span> modelscope <span class=\"token keyword\">import<\/span> AutoModelForCausalLM<span class=\"token punctuation\">,<\/span> AutoTokenizer\r\nmodel <span class=\"token operator\">=<\/span> AutoModelForCausalLM<span class=\"token punctuation\">.<\/span>from_pretrained<span class=\"token punctuation\">(<\/span><span class=\"token string\">\"qwen\/Qwen1.5-110B-Chat\"<\/span><span class=\"token punctuation\">)<\/span>\r\ntokenizer <span class=\"token operator\">=<\/span> AutoTokenizer<span class=\"token punctuation\">.<\/span>from_pretrained<span class=\"token punctuation\">(<\/span><span class=\"token string\">\"qwen\/Qwen1.5-110B-Chat\"<\/span><span class=\"token punctuation\">)<\/span>\r\n<\/code><\/pre>\n<\/li>\n<li><strong>\u51c6\u5907\u8f93\u5165<\/strong>\uff1a\u5b9a\u4e49\u63d0\u793a\uff08prompt\uff09\u548c\u7cfb\u7edf\u6d88\u606f\uff0c\u6784\u5efa\u8f93\u5165\u6587\u672c\u3002<\/li>\n<li><strong>\u751f\u6210\u6587\u672c<\/strong>\uff1a\u4f7f\u7528\u6a21\u578b\u751f\u6210\u54cd\u5e94\u6587\u672c\u3002\n<pre><code class=\"lang-python\">generated_ids <span class=\"token operator\">=<\/span> model<span class=\"token punctuation\">.<\/span>generate<span class=\"token punctuation\">(<\/span>model_inputs<span class=\"token punctuation\">.<\/span>input_ids<span class=\"token punctuation\">,<\/span> max_new_tokens<span class=\"token operator\">=<\/span><span class=\"token number\">512<\/span><span class=\"token punctuation\">)<\/span>\r\nresponse <span class=\"token operator\">=<\/span> tokenizer<span class=\"token punctuation\">.<\/span>batch_decode<span class=\"token punctuation\">(<\/span>generated_ids<span class=\"token punctuation\">,<\/span> skip_special_tokens<span class=\"token operator\">=<\/span><span class=\"token boolean\">True<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">[<\/span><span class=\"token number\">0<\/span><span class=\"token punctuation\">]<\/span>\r\n<\/code><\/pre>\n<\/li>\n<li><strong>\u8c03\u6574\u548c\u4f18\u5316<\/strong>\uff1a\u6839\u636e\u9700\u8981\u8c03\u6574\u751f\u6210\u53c2\u6570\uff0c\u5982\u00a0<code>max_new_tokens<\/code>\uff0c\u6216\u4f7f\u7528\u63d0\u4f9b\u7684\u00a0<code>generation_config.json<\/code>\u00a0\u8fdb\u884c\u8d85\u53c2\u6570\u8c03\u6574\u3002<\/li>\n<\/ol>\n<h2 id=\"-\">\u5e38\u89c1\u95ee\u9898<\/h2>\n<ul>\n<li><strong>Q:<\/strong>\u00a0\u5982\u4f55\u89e3\u51b3\u00a0<code>KeyError: 'qwen2'<\/code>\u00a0\u9519\u8bef\uff1f<br \/>\n<strong>A:<\/strong>\u00a0\u786e\u4fdd\u5df2\u5b89\u88c5\u6b63\u786e\u7248\u672c\u7684\u00a0<code>transformers<\/code>\u00a0\u5e93\u3002<\/li>\n<li><strong>Q:<\/strong>\u00a0\u5982\u4f55\u5904\u7406\u6a21\u578b\u751f\u6210\u7684\u6587\u672c\u4e0d\u7b26\u5408\u9884\u671f\uff1f<br \/>\n<strong>A:<\/strong>\u00a0\u68c0\u67e5\u8f93\u5165\u63d0\u793a\u662f\u5426\u5408\u9002\uff0c\u6216\u5c1d\u8bd5\u8c03\u6574\u751f\u6210\u53c2\u6570\u3002<\/li>\n<li><strong>Q:<\/strong>\u00a0\u5982\u4f55\u5728\u591a\u8bed\u8a00\u73af\u5883\u4e2d\u4f7f\u7528\u6a21\u578b\uff1f<br \/>\n<strong>A:<\/strong>\u00a0Qwen1.5-110B-Chat \u652f\u6301\u591a\u8bed\u8a00\uff0c\u53ea\u9700\u786e\u4fdd\u8f93\u5165\u6587\u672c\u662f\u6a21\u578b\u652f\u6301\u7684\u8bed\u8a00\u4e4b\u4e00\u3002<\/li>\n<\/ul>\n<h2 id=\"-\">\u76f8\u5173\u8d44\u6e90<\/h2>\n<ul>\n<li><strong>\u5b98\u65b9\u6587\u6863<\/strong>\uff1a<a target=\"_blank\" rel=\"noopener noreferrer nofollow\" href=\"https:\/\/www.modelscope.cn\/models\/qwen\/Qwen1.5-110B-Chat\">Qwen1.5-110B-Chat \u5b98\u65b9\u4ecb\u7ecd<\/a><\/li>\n<li><strong>GitHub \u4ed3\u5e93<\/strong>\uff1a\u8bbf\u95ee\u00a0<a target=\"_blank\" rel=\"noopener noreferrer nofollow\" href=\"https:\/\/github.com\/QwenLM\/Qwen\">Qwen GitHub \u4ed3\u5e93<\/a>\u00a0\u83b7\u53d6\u66f4\u591a\u4ee3\u7801\u793a\u4f8b\u548c\u6587\u6863\u3002<\/li>\n<li><strong>\u6280\u672f\u62a5\u544a<\/strong>\uff1a\u9605\u8bfb\u00a0<a target=\"_blank\" rel=\"noopener noreferrer nofollow\" href=\"https:\/\/qwenlm.github.io\/blog\/qwen1.5-110b\">Qwen Technical Report<\/a>\u00a0\u4ee5\u83b7\u53d6\u6280\u672f\u7ec6\u8282\u548c\u6a21\u578b\u6027\u80fd\u8bc4\u4f30\u3002<\/li>\n<\/ul>\n<h3 id=\"-\">\u5f15\u7528\u683c\u5f0f<\/h3>\n<p>\u5982\u679c\u60a8\u5728\u7814\u7a76\u4e2d\u4f7f\u7528\u4e86 Qwen1.5-110B-Chat \u6a21\u578b\uff0c\u8bf7\u6309\u7167\u4ee5\u4e0b\u683c\u5f0f\u5f15\u7528\uff1a<\/p>\n<pre><code>@article{qwen,\r\n  title={Qwen Technical Report},\r\n  author={Bai, Jinze and others},\r\n  journal={arXiv preprint arXiv:2309.16609},\r\n  year={2023}\r\n}\r\n<\/code><\/pre>\n","protected":false},"excerpt":{"rendered":"<p>Qwen1.5-110B-Chat \u662f Qwen2 \u7cfb\u5217\u7684\u6d4b\u8bd5\u7248\uff0c\u5c5e\u4e8e\u57fa\u4e8e Transformer \u67b6\u6784\u7684\u89e3\u7801\u5668\u4ec5\u8bed\u8a00\u6a21\u578b\u3002\u5b83\u7531\u4e0d\u540c\u5c3a\u5bf8\u7684\u6a21\u578b\u7ec4\u6210\uff0c\u4ece 0.5B \u5230 110B\uff0c\u5305\u62ec\u4e00\u4e2a 14B \u7684 MoE \u6a21\u578b\uff0c\u5177\u6709 2.7B \u7684\u6fc0\u6d3b\u91cf\u3002\u8be5\u6a21\u578b\u5728\u4eba\u7c7b\u504f\u597d\u3001\u591a\u8bed\u8a00\u652f\u6301\u3001\u4e0a\u4e0b\u6587\u957f\u5ea6\u7a33\u5b9a\u6027\u7b49\u65b9\u9762\u8fdb\u884c\u4e86\u663e\u8457\u6539\u8fdb\uff0c\u5e76\u652f\u6301\u9ad8\u8fbe 32K \u7684\u4e0a\u4e0b\u6587\u957f\u5ea6\u3002 \u6a21\u578b\u8bc4\u6d4b Qwen1.5-110B-Chat \u5728\u591a\u4e2a\u7ef4\u5ea6\u4e0a\u8fdb\u884c\u4e86\u6027\u80fd\u8bc4\u6d4b\uff0c\u5305\u62ec\u4f46\u4e0d\u9650\u4e8e\u8bed\u8a00\u751f\u6210\u80fd\u529b\u3001\u591a\u8bed\u8a00\u7406\u89e3\u548c\u751f\u6210\u3001\u4ee5\u53ca\u5728\u7279\u5b9a\u4efb\u52a1\u4e0a\u7684\u8868\u73b0\u3002\u6a21\u578b\u7684\u8bc4\u6d4b\u7ed3\u679c\u8868\u660e\uff0c\u5b83\u5728\u804a\u5929\u6a21\u578b\u7684\u4eba<\/p>\n","protected":false},"author":1,"featured_media":210,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3],"tags":[],"collection":[42],"company":[6],"rank":[54,53],"class_list":["post-209","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-llm","collection-qwen1-5","company-alibaba","rank-llm-leaderboard","rank-china-llm-ranking"],"_links":{"self":[{"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/posts\/209","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=209"}],"version-history":[{"count":2,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/posts\/209\/revisions"}],"predecessor-version":[{"id":212,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/posts\/209\/revisions\/212"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/media\/210"}],"wp:attachment":[{"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/media?parent=209"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/categories?post=209"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/tags?post=209"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/collection?post=209"},{"taxonomy":"company","embeddable":true,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/company?post=209"},{"taxonomy":"rank","embeddable":true,"href":"https:\/\/www.onetts.com\/ai\/wp-json\/wp\/v2\/rank?post=209"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}