{"id":32899,"date":"2024-09-30T00:12:06","date_gmt":"2024-09-29T16:12:06","guid":{"rendered":"https:\/\/17aitech.com\/?p=32899"},"modified":"2024-11-30T16:37:36","modified_gmt":"2024-11-30T08:37:36","slug":"%e3%80%90%e8%af%be%e7%a8%8b%e6%80%bb%e7%bb%93%e3%80%91day31%ef%bc%9a%e5%a4%9a%e6%a8%a1%e6%80%81%e5%a4%a7%e6%a8%a1%e5%9e%8b%e5%88%9d%e6%ad%a5%e4%ba%86%e8%a7%a3","status":"publish","type":"post","link":"https:\/\/17aitech.com\/?p=32899","title":{"rendered":"\u3010\u8bfe\u7a0b\u603b\u7ed3\u3011day31\uff1a\u591a\u6a21\u6001\u5927\u6a21\u578b\u521d\u6b65\u4e86\u89e3"},"content":{"rendered":"<div id=\"ez-toc-container\" class=\"ez-toc-v2_0_78 ez-toc-wrap-left-text counter-hierarchy ez-toc-counter ez-toc-light-blue ez-toc-container-direction\">\n<div class=\"ez-toc-title-container\">\n<p class=\"ez-toc-title\" style=\"cursor:inherit\">\u6587\u7ae0\u76ee\u5f55<\/p>\n<span class=\"ez-toc-title-toggle\"><a href=\"#\" class=\"ez-toc-pull-right ez-toc-btn ez-toc-btn-xs ez-toc-btn-default ez-toc-toggle\" aria-label=\"Toggle Table of Content\"><span class=\"ez-toc-js-icon-con\"><span class=\"\"><span class=\"eztoc-hide\" style=\"display:none;\">Toggle<\/span><span class=\"ez-toc-icon-toggle-span\"><svg style=\"fill: #999;color:#999\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" class=\"list-377408\" width=\"20px\" height=\"20px\" viewBox=\"0 0 24 24\" fill=\"none\"><path d=\"M6 6H4v2h2V6zm14 0H8v2h12V6zM4 11h2v2H4v-2zm16 0H8v2h12v-2zM4 16h2v2H4v-2zm16 0H8v2h12v-2z\" fill=\"currentColor\"><\/path><\/svg><svg style=\"fill: #999;color:#999\" class=\"arrow-unsorted-368013\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" width=\"10px\" height=\"10px\" viewBox=\"0 0 24 24\" version=\"1.2\" baseProfile=\"tiny\"><path d=\"M18.2 9.3l-6.2-6.3-6.2 6.3c-.2.2-.3.4-.3.7s.1.5.3.7c.2.2.4.3.7.3h11c.3 0 .5-.1.7-.3.2-.2.3-.5.3-.7s-.1-.5-.3-.7zM5.8 14.7l6.2 6.3 6.2-6.3c.2-.2.3-.5.3-.7s-.1-.5-.3-.7c-.2-.2-.4-.3-.7-.3h-11c-.3 0-.5.1-.7.3-.2.2-.3.5-.3.7s.1.5.3.7z\"\/><\/svg><\/span><\/span><\/span><\/a><\/span><\/div>\n<nav><ul class='ez-toc-list ez-toc-list-level-1 ' ><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-1\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%89%8D%E8%A8%80\" >\u524d\u8a00<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-2\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%A4%A7%E8%A7%86%E8%A7%89%E6%A8%A1%E5%9E%8B%E7%9A%84%E4%BD%93%E9%AA%8C%E6%84%9F%E5%8F%97\" >\u5927\u89c6\u89c9\u6a21\u578b\u7684\u4f53\u9a8c\u611f\u53d7<\/a><ul class='ez-toc-list-level-3' ><li class='ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-3\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%9C%BA%E6%99%AF%E4%B8%80%EF%BC%9A%E5%9B%BE%E5%83%8F%E8%AF%86%E5%88%AB\" >\u573a\u666f\u4e00\uff1a\u56fe\u50cf\u8bc6\u522b<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-4\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%9C%BA%E6%99%AF%E4%BA%8C%EF%BC%9AOCR%E6%96%87%E5%AD%97%E8%AF%86%E5%88%AB%E5%9F%BA%E6%9C%AC%E6%96%87%E5%AD%97\" >\u573a\u666f\u4e8c\uff1aOCR\u6587\u5b57\u8bc6\u522b(\u57fa\u672c\u6587\u5b57)<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-5\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%9C%BA%E6%99%AF%E4%B8%89%EF%BC%9AOCR%E6%96%87%E5%AD%97%E8%AF%86%E5%88%AB%E6%8E%A8%E7%90%86\" >\u573a\u666f\u4e09\uff1aOCR\u6587\u5b57\u8bc6\u522b+\u63a8\u7406<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-6\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%9C%BA%E6%99%AF%E5%9B%9B%EF%BC%9AOCR%E6%96%87%E5%AD%97%E8%AF%86%E5%88%AB%E4%BB%A3%E7%A0%81%E8%AF%86%E5%88%AB%E5%B9%B6%E6%94%B9%E9%94%99\" >\u573a\u666f\u56db\uff1aOCR\u6587\u5b57\u8bc6\u522b(\u4ee3\u7801\u8bc6\u522b\u5e76\u6539\u9519)<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-7\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%9C%BA%E6%99%AF%E4%BA%94%EF%BC%9A%E7%9B%AE%E6%A0%87%E6%A3%80%E6%B5%8B\" >\u573a\u666f\u4e94\uff1a\u76ee\u6807\u68c0\u6d4b<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-8\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%B0%8F%E7%BB%93\" >\u5c0f\u7ed3<\/a><\/li><\/ul><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-9\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%A4%9A%E6%A8%A1%E6%80%81%E5%A4%A7%E6%A8%A1%E5%9E%8B%E4%BB%8B%E7%BB%8D\" >\u591a\u6a21\u6001\u5927\u6a21\u578b\u4ecb\u7ecd<\/a><ul class='ez-toc-list-level-3' ><li class='ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-10\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E7%AE%80%E4%BB%8B\" >\u7b80\u4ecb<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-11\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%8F%91%E5%B1%95%E5%8E%86%E5%8F%B2\" >\u53d1\u5c55\u5386\u53f2<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-12\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%A4%9A%E6%A8%A1%E6%80%81%E5%A4%A7%E6%A8%A1%E5%9E%8B%E6%9E%B6%E6%9E%84%E7%BB%84%E6%88%90\" >\u591a\u6a21\u6001\u5927\u6a21\u578b\u67b6\u6784\u7ec4\u6210<\/a><ul class='ez-toc-list-level-4' ><li class='ez-toc-heading-level-4'><a class=\"ez-toc-link ez-toc-heading-13\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E6%A8%A1%E6%80%81%E7%BC%96%E7%A0%81%E5%99%A8%E5%9B%BE%E2%91%A0\" >\u6a21\u6001\u7f16\u7801\u5668(\u56fe\u2460)<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-4'><a class=\"ez-toc-link ez-toc-heading-14\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E6%A8%A1%E6%80%81%E8%BF%9E%E6%8E%A5%E5%99%A8%E5%9B%BE%E2%91%A1\" >\u6a21\u6001\u8fde\u63a5\u5668(\u56fe\u2461)<\/a><ul class='ez-toc-list-level-5' ><li class='ez-toc-heading-level-5'><a class=\"ez-toc-link ez-toc-heading-15\" href=\"https:\/\/17aitech.com\/?p=32899\/#MLP_%E5%9F%BA%E4%BA%8E%E6%8A%95%E5%BD%B1%E7%9A%84%E8%BF%9E%E6%8E%A5%E5%99%A8%E5%9B%BE%E2%91%A4\" >MLP \u57fa\u4e8e\u6295\u5f71\u7684\u8fde\u63a5\u5668(\u56fe\u2464)<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-5'><a class=\"ez-toc-link ez-toc-heading-16\" href=\"https:\/\/17aitech.com\/?p=32899\/#Q-Former_%E5%9F%BA%E4%BA%8E%E6%9F%A5%E8%AF%A2%E7%9A%84%E8%BF%9E%E6%8E%A5%E5%99%A8%E5%9B%BE%E2%91%A5\" >Q-Former \u57fa\u4e8e\u67e5\u8be2\u7684\u8fde\u63a5\u5668(\u56fe\u2465)<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-5'><a class=\"ez-toc-link ez-toc-heading-17\" href=\"https:\/\/17aitech.com\/?p=32899\/#MH-Attn_%E5%9F%BA%E4%BA%8E%E8%9E%8D%E5%90%88%E7%9A%84%E8%BF%9E%E6%8E%A5%E5%99%A8%E5%9B%BE%E2%91%A6\" >MH-Attn \u57fa\u4e8e\u878d\u5408\u7684\u8fde\u63a5\u5668(\u56fe\u2466)<\/a><\/li><\/ul><\/li><li class='ez-toc-page-1 ez-toc-heading-level-4'><a class=\"ez-toc-link ez-toc-heading-18\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E6%A8%A1%E6%80%81%E7%94%9F%E6%88%90%E5%99%A8%E5%9B%BE%E2%91%A3\" >\u6a21\u6001\u751f\u6210\u5668(\u56fe\u2463)<\/a><\/li><\/ul><\/li><\/ul><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-19\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%A4%9A%E6%A8%A1%E6%80%81%E5%A4%A7%E6%A8%A1%E5%9E%8B%E8%AE%AD%E7%BB%83\" >\u591a\u6a21\u6001\u5927\u6a21\u578b\u8bad\u7ec3<\/a><ul class='ez-toc-list-level-3' ><li class='ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-20\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E7%AC%AC%E4%B8%80%E9%98%B6%E6%AE%B5%EF%BC%9A%E9%A2%84%E8%AE%AD%E7%BB%83Pre-train\" >\u7b2c\u4e00\u9636\u6bb5\uff1a\u9884\u8bad\u7ec3(Pre-train)<\/a><ul class='ez-toc-list-level-4' ><li class='ez-toc-heading-level-4'><a class=\"ez-toc-link ez-toc-heading-21\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E7%B2%97%E7%B2%92%E5%BA%A6%E5%9B%BE%E5%83%8F-%E6%96%87%E5%AD%97%E6%95%B0%E6%8D%AE%E9%9B%86%E4%B8%80%E8%A7%88\" >\u7c97\u7c92\u5ea6\u56fe\u50cf-\u6587\u5b57\u6570\u636e\u96c6\u4e00\u89c8<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-4'><a class=\"ez-toc-link ez-toc-heading-22\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E7%BB%86%E7%B2%92%E5%BA%A6%E5%9B%BE%E5%83%8F-%E6%96%87%E5%AD%97%E6%95%B0%E6%8D%AE%E9%9B%86%E4%B8%80%E8%A7%88\" >\u7ec6\u7c92\u5ea6\u56fe\u50cf-\u6587\u5b57\u6570\u636e\u96c6\u4e00\u89c8<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-4'><a class=\"ez-toc-link ez-toc-heading-23\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E8%A7%86%E9%A2%91-%E6%96%87%E5%AD%97%E6%95%B0%E6%8D%AE%E9%9B%86%E4%B8%80%E8%A7%88\" >\u89c6\u9891-\u6587\u5b57\u6570\u636e\u96c6\u4e00\u89c8<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-4'><a class=\"ez-toc-link ez-toc-heading-24\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E9%9F%B3%E9%A2%91-%E6%96%87%E5%AD%97%E6%95%B0%E6%8D%AE%E9%9B%86%E4%B8%80%E8%A7%88\" >\u97f3\u9891-\u6587\u5b57\u6570\u636e\u96c6\u4e00\u89c8<\/a><\/li><\/ul><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-25\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E7%AC%AC%E4%BA%8C%E9%98%B6%E6%AE%B5%EF%BC%9A%E5%BE%AE%E8%B0%83fine-tune\" >\u7b2c\u4e8c\u9636\u6bb5\uff1a\u5fae\u8c03(fine-tune)<\/a><ul class='ez-toc-list-level-4' ><li class='ez-toc-heading-level-4'><a class=\"ez-toc-link ez-toc-heading-26\" href=\"https:\/\/17aitech.com\/?p=32899\/#LLaVa-Instruct\" >LLaVa-Instruct<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-4'><a class=\"ez-toc-link ez-toc-heading-27\" href=\"https:\/\/17aitech.com\/?p=32899\/#Video-ChatGPT\" >Video-ChatGPT<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-4'><a class=\"ez-toc-link ez-toc-heading-28\" href=\"https:\/\/17aitech.com\/?p=32899\/#Clotho\" >Clotho<\/a><\/li><\/ul><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-29\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E7%AC%AC%E4%B8%89%E9%98%B6%E6%AE%B5%EF%BC%9A%E5%81%8F%E5%A5%BD%E5%AF%B9%E9%BD%90\" >\u7b2c\u4e09\u9636\u6bb5\uff1a\u504f\u597d\u5bf9\u9f50<\/a><\/li><\/ul><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-30\" href=\"https:\/\/17aitech.com\/?p=32899\/#Qwen2%E7%9A%84%E9%83%A8%E7%BD%B2%E5%AE%9E%E8%B7%B5\" >Qwen2\u7684\u90e8\u7f72\u5b9e\u8df5<\/a><ul class='ez-toc-list-level-3' ><li class='ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-31\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%90%AF%E5%8A%A8%E7%8E%AF%E5%A2%83\" >\u542f\u52a8\u73af\u5883<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-32\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E6%A8%A1%E5%9E%8B%E4%B8%8B%E8%BD%BD\" >\u6a21\u578b\u4e0b\u8f7d<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-33\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%AE%89%E8%A3%85%E4%BE%9D%E8%B5%96\" >\u5b89\u88c5\u4f9d\u8d56<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-34\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E6%A8%A1%E5%9E%8B%E5%8A%A0%E8%BD%BD\" >\u6a21\u578b\u52a0\u8f7d<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-35\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E6%9E%84%E5%BB%BAprompt\" >\u6784\u5efaprompt<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-36\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%87%86%E5%A4%87%E6%8E%A8%E7%90%86%E6%95%B0%E6%8D%AE\" >\u51c6\u5907\u63a8\u7406\u6570\u636e<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-37\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E6%A8%A1%E5%9E%8B%E6%8E%A8%E7%90%86\" >\u6a21\u578b\u63a8\u7406<\/a><\/li><\/ul><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-38\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E9%87%8F%E5%8C%96%E5%86%85%E5%AE%B9%E7%AE%80%E4%BB%8B\" >\u91cf\u5316\u5185\u5bb9\u7b80\u4ecb<\/a><ul class='ez-toc-list-level-3' ><li class='ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-39\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E9%87%8F%E5%8C%96%E7%9A%84%E5%AE%9A%E4%B9%89\" >\u91cf\u5316\u7684\u5b9a\u4e49<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-40\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E9%87%8F%E5%8C%96%E7%9A%84%E8%83%8C%E6%99%AF\" >\u91cf\u5316\u7684\u80cc\u666f<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-41\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E9%87%8F%E5%8C%96%E7%9A%84%E5%8E%9F%E7%90%86\" >\u91cf\u5316\u7684\u539f\u7406<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-42\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E9%87%8F%E5%8C%96%E7%9A%84%E4%BC%98%E7%82%B9\" >\u91cf\u5316\u7684\u4f18\u70b9<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-43\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E9%87%8F%E5%8C%96%E7%9A%84%E6%8C%91%E6%88%98\" >\u91cf\u5316\u7684\u6311\u6218<\/a><\/li><\/ul><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-44\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%86%85%E5%AE%B9%E5%B0%8F%E7%BB%93\" >\u5185\u5bb9\u5c0f\u7ed3<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-45\" href=\"https:\/\/17aitech.com\/?p=32899\/#%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99\" >\u53c2\u8003\u8d44\u6599<\/a><\/li><\/ul><\/nav><\/div>\n<h2><span class=\"ez-toc-section\" id=\"%E5%89%8D%E8%A8%80\"><\/span>\u524d\u8a00<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u968f\u77402024\u5e745\u670814\u65e5GPT-4o\u7684\u53d1\u5e03\uff0c\u591a\u6a21\u6001\u5927\u6a21\u578b\u9010\u6e10\u6210\u4e3a\u884c\u4e1a\u70ed\u70b9\uff0c\u56fd\u5185\u5404\u5927\u5382\u5546\u963f\u91cc\u3001\u767e\u5ea6\u3001\u667a\u8c31\u4e5f\u76f8\u7ee7\u53d1\u5e03\u4e86\u81ea\u5df1\u7684\u591a\u6a21\u6001\u5927\u6a21\u578b\u3002\u672c\u7ae0\u5185\u5bb9\uff0c\u6211\u4eec\u5c06\u4ece\u591a\u6a21\u6001\u7684\u4f53\u9a8c\u611f\u53d7\u5f00\u59cb\uff0c\u9010\u6b65\u4e86\u89e3\u591a\u6a21\u6001\u5927\u6a21\u578b\u7684\u67b6\u6784\u7ec4\u6210\u3001\u8bad\u7ec3\u6570\u636e\u6784\u6210\uff0c\u6700\u540e\u642d\u5efa\u4e00\u4e2a\u591a\u6a21\u6001\u5927\u6a21\u578b\u3002<\/p>\n<h2><span class=\"ez-toc-section\" id=\"%E5%A4%A7%E8%A7%86%E8%A7%89%E6%A8%A1%E5%9E%8B%E7%9A%84%E4%BD%93%E9%AA%8C%E6%84%9F%E5%8F%97\"><\/span>\u5927\u89c6\u89c9\u6a21\u578b\u7684\u4f53\u9a8c\u611f\u53d7<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u4e3a\u4e86\u5bf9\u5927\u6a21\u578b\u6709\u4e00\u4e2a\u521d\u6b65\u7684\u4e86\u89e3\uff0c\u6211\u4eec\u4f7f\u7528\u963f\u91cc\u6700\u65b0\u516c\u5e03\u7684\u901a\u4e49\u5343\u95eeVL-Max\uff0c\u611f\u53d7\u4e00\u4e0b\u591a\u6a21\u6001\u5927\u6a21\u578b\u7684\u80fd\u529b\u3002<\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E5%9C%BA%E6%99%AF%E4%B8%80%EF%BC%9A%E5%9B%BE%E5%83%8F%E8%AF%86%E5%88%AB\"><\/span>\u573a\u666f\u4e00\uff1a\u56fe\u50cf\u8bc6\u522b<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><strong>\u9996\u5148<\/strong>\uff0c\u6211\u4eec\u641c\u7d22\u4e00\u4e2a\u5173\u4e8e\u5bfc\u5f39\u7684\u56fe\u7247(\u76f8\u4fe1\u6709\u5f88\u591a\u4eba\u8ddf\u6211\u4e00\u6837\uff0c\u5bf9\u6211\u56fd\u56fd\u9632\u529b\u91cf\u611f\u5174\u8da3\uff0c\u4f46\u662f\u4e0d\u77e5\u9053\u5177\u4f53\u7684\u6b66\u5668\u578b\u53f7)\u3002<\/p>\n<p><strong>\u63a5\u7740<\/strong>\uff0c\u6211\u4eec\u5728<a href=\"https:\/\/bailian.console.aliyun.com\/\">\u963f\u91cc\u4e91\u767e\u70bc<\/a>\u5e73\u53f0\uff0c\u9009\u62e9\u901a\u4e49\u5343\u95eeVL-Max\u6a21\u578b\uff0c\u4e0a\u4f20\u56fe\u7247\u5e76\u63d0\u95ee\u3002\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u4e1c\u98ce21D.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u4e1c\u98ce21D.png\" alt=\"\" \/><\/a><\/p>\n<p>\u5982\u4e0a\u56fe\u6240\u793a\uff0c\u591a\u6a21\u6001\u5927\u6a21\u578b\u5f88\u5feb\u5c31\u8bc6\u522b\u51fa\u4e86\u8be5\u5bfc\u5f39\u7684\u5177\u4f53\u578b\u53f7\u5e76\u4e14\u7ed9\u51fa\u4e86\u76f8\u5173\u7684\u4ecb\u7ecd\u3002<\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E5%9C%BA%E6%99%AF%E4%BA%8C%EF%BC%9AOCR%E6%96%87%E5%AD%97%E8%AF%86%E5%88%AB%E5%9F%BA%E6%9C%AC%E6%96%87%E5%AD%97\"><\/span>\u573a\u666f\u4e8c\uff1aOCR\u6587\u5b57\u8bc6\u522b(\u57fa\u672c\u6587\u5b57)<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>OCR\u6587\u5b57\u8bc6\u522b\u672c\u8eab\u5176\u5b9e\u662f\u4e00\u9879\u975e\u5e38\u6210\u719f\u7684\u6280\u672f\u4e86\uff0c\u4f46\u662f\u5728\u5927\u6a21\u578b\u7684\u52a0\u6301\u4e0b\uff0c\u6211\u4eec\u770b\u770b\u5176\u80fd\u529b\u5982\u4f55\u3002<\/p>\n<p><strong>\u9996\u5148<\/strong>\uff0c\u6211\u4eec\u5728\u641c\u7d22\u5f15\u64ce\u91cc\u968f\u610f\u641c\u7d22\u4e00\u5f20\u706b\u8f66\u7968\u3002<\/p>\n<p><strong>\u63a5\u7740<\/strong>\uff0c\u6211\u4eec\u5728\u963f\u91cc\u4e91\u767e\u70bc\u5e73\u53f0\u4e0a\uff0c\u4e0a\u4f20\u56fe\u7247\u5e76\u63d0\u95ee\uff0c\u63d0\u95ee\u5185\u5bb9\u6211\u4eec\u4e0d\u4ec5\u9650\u4e8e\u6587\u5b57\u8bc6\u522b\uff0c\u800c\u662f\u589e\u52a0\u4e00\u4e9b\u63a8\u7406\u5185\u5bb9\uff0c\u4f8b\u5982\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u706b\u8f66\u7968\u8bc6\u522b.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u706b\u8f66\u7968\u8bc6\u522b.png\" alt=\"\" \/><\/a><\/p>\n<p>\u53ef\u4ee5\u770b\u5230\uff0c\u591a\u6a21\u6001\u5927\u6a21\u578b\u8f7b\u6613\u5730\u5c31\u8bc6\u522b\u51fa\u4e86\u56fe\u7247\u4e2d\u706b\u8f66\u7968\u7684\u91d1\u989d\uff0c\u5e76\u4e14\u6309\u7167\u9898\u76ee\u8981\u6c42\u81ea\u52a8\u8ba1\u7b97\u4e86\u603b\u5171\u82b1\u8d39\u7684\u91d1\u989d\u3002<\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E5%9C%BA%E6%99%AF%E4%B8%89%EF%BC%9AOCR%E6%96%87%E5%AD%97%E8%AF%86%E5%88%AB%E6%8E%A8%E7%90%86\"><\/span>\u573a\u666f\u4e09\uff1aOCR\u6587\u5b57\u8bc6\u522b+\u63a8\u7406<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u573a\u666f\u4e8c\u662f\u4e00\u4e2a\u7b80\u5355\u7684OCR\u8bc6\u522b\uff0c\u63a5\u4e0b\u6765\u6211\u4eec\u8bd5\u4e00\u4e2aOCR\u8bc6\u522b+\u63a8\u7406\u7684\u4f8b\u5b50\u3002<\/p>\n<p><strong>\u9996\u5148<\/strong>\uff0c\u6211\u4eec\u5728\u641c\u7d22\u5f15\u64ce\u91cc\u641c\u7d22\u4e00\u5f20\u533b\u9662\u7684\u5bfc\u8bca\u7167\u7247\uff0c\u5982\u56fe\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u5bfc\u8bca\u724c\u5b50.jpg\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u5bfc\u8bca\u724c\u5b50.jpg\" alt=\"\" \/><\/a><\/p>\n<p><strong>\u63a5\u7740<\/strong>\uff0c\u6211\u4eec\u4e0a\u4f20\u56fe\u7247\u5e76\u63d0\u95ee\u4e00\u4e2a\u5e26\u6709\u63a8\u7406\u6027\u7684\u95ee\u9898\uff0c\u4f8b\u5982\uff1a\u201c\u6211\u611f\u5192\u4e86\uff0c\u8bf7\u6839\u636e\u56fe\u7247\u63d0\u793a\uff0c\u544a\u8bc9\u6211\u5177\u4f53\u770b\u75c5\u662f\u600e\u6837\u7684\u6d41\u7a0b\u3002\u201d\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u5bfc\u8bca\u8bc6\u522b\u63a8\u7406.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u5bfc\u8bca\u8bc6\u522b\u63a8\u7406.png\" alt=\"\" \/><\/a><\/p>\n<p>\u901a\u8fc7\u4e0a\u56fe\u53ef\u4ee5\u770b\u5230\uff0c\u591a\u6a21\u6001\u5927\u6a21\u578b\u4e0d\u4ec5\u51c6\u786e\u8bc6\u522b\u51fa\u56fe\u7247\u4e2d\u5404\u4e2a\u79d1\u5ba4\u7684\u4f4d\u7f6e\uff0c\u8fd8\u501f\u52a9\u5927\u6a21\u578b\u7684\u63a8\u7406\u80fd\u529b\uff0c\u9488\u5bf9\u6211\u7684\u63d0\u95ee\u7ed9\u51fa\u4e86\u76f8\u5e94\u7684\u6d41\u7a0b\u3002<\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E5%9C%BA%E6%99%AF%E5%9B%9B%EF%BC%9AOCR%E6%96%87%E5%AD%97%E8%AF%86%E5%88%AB%E4%BB%A3%E7%A0%81%E8%AF%86%E5%88%AB%E5%B9%B6%E6%94%B9%E9%94%99\"><\/span>\u573a\u666f\u56db\uff1aOCR\u6587\u5b57\u8bc6\u522b(\u4ee3\u7801\u8bc6\u522b\u5e76\u6539\u9519)<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u63a5\u4e0b\u6765\uff0c\u6211\u4eec\u8bd5\u9a8c\u4e00\u4e2a\u591a\u6a21\u6001\u5927\u6a21\u578b\u8bc6\u522b\u4ee3\u7801\u9519\u8bef\u5e76\u4fee\u6539\u7684\u793a\u4f8b\u3002<\/p>\n<p><strong>\u9996\u5148<\/strong>\uff0c\u6211\u4eec\u5bf9\u4e00\u6bb5\u6709\u9519\u8bef\u7684\u4ee3\u7801\u8fdb\u884c\u622a\u56fe\uff0c\u622a\u56fe\u5185\u5bb9\u5982\u4e0b\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u4ee3\u7801\u6539\u9519.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u4ee3\u7801\u6539\u9519.png\" alt=\"\" \/><\/a><\/p>\n<p><strong>\u63a5\u7740<\/strong>\uff0c\u6211\u4eec\u4e0a\u4f20\u56fe\u7247\u7ed9\u591a\u6a21\u6001\u5927\u6a21\u578b\u5e76\u4e14\u8ba9\u5927\u6a21\u578b\u627e\u5230\u9519\u8bef\u5e76\u4fee\u6539\u3002\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u4ee3\u7801\u6539\u9519\u7ed3\u679c.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u4ee3\u7801\u6539\u9519\u7ed3\u679c.png\" alt=\"\" \/><\/a><\/p>\n<p>\u901a\u8fc7\u4e0a\u56fe\u53ef\u4ee5\u770b\u5230\uff0c\u591a\u6a21\u6001\u5927\u6a21\u578b\u51c6\u786e\u5730\u8bc6\u522b\u51fa\u4ee3\u7801\u4e2d\u7684\u9519\u8bef\u5e76\u4e14\u7ed9\u51fa\u4e86\u6b63\u786e\u7684\u4ee3\u7801\u7247\u6bb5\u3002<\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E5%9C%BA%E6%99%AF%E4%BA%94%EF%BC%9A%E7%9B%AE%E6%A0%87%E6%A3%80%E6%B5%8B\"><\/span>\u573a\u666f\u4e94\uff1a\u76ee\u6807\u68c0\u6d4b<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u5728<a href=\"https:\/\/17aitech.com\/?p=2398\">\u3010\u8bfe\u7a0b\u603b\u7ed3\u3011Day13\uff08\u4e0a\uff09\uff1a\u4f7f\u7528YOLO\u8fdb\u884c\u76ee\u6807\u68c0\u6d4b<\/a>\u5b66\u4e60\u4e2d\uff0c\u6211\u4eec\u5bf9\u56fe\u7247\u4e2d\u7684\u76ee\u6807\u8fdb\u884c\u68c0\u6d4b\u7684\u8bdd\uff0c\u4e00\u822c\u9700\u8981\u51c6\u5907\u5927\u91cf\u7684\u6570\u636e\u7136\u540e\u501f\u52a9YOLO\u8bad\u7ec3\uff0c\u624d\u53ef\u4ee5\u8bc6\u522b\u51fa\u76ee\u6807\u3002\u90a3\u4e48\uff0c\u5982\u679c\u662f\u591a\u6a21\u6001\u5927\u6a21\u578b\u4f1a\u600e\u6837\u5462\uff1a<\/p>\n<p><strong>\u9996\u5148<\/strong>\uff0c\u6211\u4eec\u641c\u7d22\u4e00\u5f20&quot;\u4e03\u9f99\u73e0\u7684\u5168\u5bb6\u798f\u56fe\u7247&quot;\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u4e03\u9f99\u73e0\u5168\u5bb6\u798f.jpg\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u4e03\u9f99\u73e0\u5168\u5bb6\u798f.jpg\" alt=\"\" \/><\/a><\/p>\n<p><strong>\u63a5\u7740<\/strong>\uff0c\u6211\u4eec\u4e0a\u4f20\u56fe\u7247\u7ed9\u591a\u6a21\u6001\u5927\u6a21\u578b\u5e76\u4e14\u8ba9\u5176\u5708\u51fa\u4eba\u9020\u4eba18\u53f7\u7684\u4f4d\u7f6e\u3002\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u8bc6\u522b\u4eba\u9020\u4eba.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u8bc6\u522b\u4eba\u9020\u4eba.png\" alt=\"\" \/><\/a><\/p>\n<p><strong>\u7136\u540e<\/strong>\uff0c\u6211\u4eec\u4f7f\u7528Python\u4ee3\u7801\u5bf9\u8bc6\u522b\u7ed3\u679c\u8fdb\u884c\u7ed8\u56fe\u3002<\/p>\n<blockquote>\n<p>\u76ee\u524d\u4f7f\u7528\u7684\u901a\u4e49\u5343\u95ee\u591a\u6a21\u6001\u5927\u6a21\u578b\u8fd8\u65e0\u6cd5\u8f93\u51fa\u56fe\u7247\uff0c\u6240\u4ee5\u6211\u4eec\u6b64\u5904\u901a\u8fc7Python\u4ee3\u7801\u624b\u52a8\u7ed8\u5236\u3002<\/p>\n<\/blockquote>\n<pre><code class=\"language-python\">from PIL import Image, ImageDraw, ImageFont\n\n# \u6253\u5f00\u56fe\u7247\nimage_path = &#039;\u4e03\u9f99\u73e0\u5168\u5bb6\u798f.jpg&#039;  # \u66ff\u6362\u4e3a\u4f60\u7684\u56fe\u7247\u8def\u5f84\nimage = Image.open(image_path)\n\n# \u521b\u5efa\u4e00\u4e2a\u53ef\u7ed8\u5236\u7684\u5bf9\u8c61\ndraw = ImageDraw.Draw(image)\n\n# \u5b9a\u4e49\u591a\u4e2a\u77e9\u5f62\u6846\u7684\u5750\u6807\u548c\u5bf9\u5e94\u7684\u4eba\u7269\u540d\u79f0\nboxes = [\n    ((20, 200, 100, 970), &quot;\u4eba\u9020\u4eba18\u53f7&quot;),\n]\n\n# \u7ed8\u5236\u77e9\u5f62\u6846\nfor (x1, y1, x2, y2), name in boxes:\n    # \u7ed8\u5236\u77e9\u5f62\u6846\n    draw.rectangle([x1, y1, x2, y2], outline=&quot;red&quot;, width=3)\n    # \u6807\u6ce8\u540d\u79f0\n    draw.text((x1, y1), name, fill=&quot;red&quot;)\n\n# \u4fdd\u5b58\u6216\u663e\u793a\u56fe\u7247\nimage.show()  # \u663e\u793a\u56fe\u7247<\/code><\/pre>\n<p>\u8fd0\u884c\u7ed3\u679c\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u4e03\u9f99\u73e0\u8bc6\u522b\u7ed3\u679c.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u4e03\u9f99\u73e0\u8bc6\u522b\u7ed3\u679c.png\" alt=\"\" \/><\/a><\/p>\n<p>\u5728\u6ca1\u6709\u4efb\u4f55\u8bad\u7ec3\u7684\u60c5\u51b5\u4e0b\uff0c\u6211\u4eec\u53ea\u662f\u63d0\u4f9b\u56fe\u7247\u7ed9\u591a\u6a21\u6001\u5927\u6a21\u578b\u5e76\u7ed9\u51fa\u6211\u4eec\u7684\u8981\u6c42\uff0c\u5b83\u5c31\u80fd\u8f83\u4e3a\u51c6\u786e\u5730\u8bc6\u522b\u51fa\u56fe\u7247\u4e2d\u4eba\u7269\u7684\u4f4d\u7f6e\u3002<\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E5%B0%8F%E7%BB%93\"><\/span>\u5c0f\u7ed3<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u901a\u8fc7\u4ee5\u4e0a\u591a\u4e2a\u573a\u666f\u7684\u8bd5\u7528\u4f53\u9a8c\uff0c\u6211\u4eec\u53ef\u4ee5\u770b\u5230\uff1a\u5927\u89c6\u89c9\u6a21\u578b\u62e5\u6709\u770b\u61c2\u56fe\u50cf\u7684\u80fd\u529b\uff0c\u5927\u8bed\u8a00\u6a21\u578b\u62e5\u6709\u5f3a\u5927\u63a8\u7406\u80fd\u529b\uff0c\u5c06\u8fd9\u4e24\u8005\u76f8\u7ed3\u5408\u7684\u591a\u6a21\u6001\u5927\u6a21\u578b\uff0c\u53ef\u4ee5\u5f00\u8f9f\u4e00\u4e2a\u65b0\u7684\u9886\u57df\u3002<\/p>\n<h2><span class=\"ez-toc-section\" id=\"%E5%A4%9A%E6%A8%A1%E6%80%81%E5%A4%A7%E6%A8%A1%E5%9E%8B%E4%BB%8B%E7%BB%8D\"><\/span>\u591a\u6a21\u6001\u5927\u6a21\u578b\u4ecb\u7ecd<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<h3><span class=\"ez-toc-section\" id=\"%E7%AE%80%E4%BB%8B\"><\/span>\u7b80\u4ecb<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u591a\u6a21\u6001\u5927\u6a21\u578b\u662f\u4e00\u79cd\u80fd\u591f\u5904\u7406\u548c\u7406\u89e3\u591a\u79cd\u6570\u636e\u7c7b\u578b\uff08\u6a21\u6001\uff09\u7684\u4eba\u5de5\u667a\u80fd\u6a21\u578b\u3002\u8fd9\u4e9b\u6a21\u6001\u901a\u5e38\u5305\u62ec\u6587\u672c\u3001\u56fe\u50cf\u3001\u97f3\u9891\u548c\u89c6\u9891\u7b49\u3002<\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E5%8F%91%E5%B1%95%E5%8E%86%E5%8F%B2\"><\/span>\u53d1\u5c55\u5386\u53f2<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><strong>1\u3001\u65e9\u671f\u9636\u6bb5\uff082010\u5e74\u4ee3\u521d\uff09<\/strong><\/p>\n<ul>\n<li>\u591a\u6a21\u6001\u5b66\u4e60\u7684\u7814\u7a76\u8d77\u6b65\u4e8e\u5bf9\u4e0d\u540c\u6a21\u6001\u6570\u636e\u7684\u72ec\u7acb\u5904\u7406\uff0c\u5982\u56fe\u50cf\u5206\u7c7b\u548c\u6587\u672c\u5904\u7406\u3002<\/li>\n<li>\u7814\u7a76\u8005\u4eec\u5f00\u59cb\u63a2\u7d22\u5982\u4f55\u5c06\u6587\u672c\u4e0e\u56fe\u50cf\u7ed3\u5408\uff0c\u4ee5\u63d0\u9ad8\u4efb\u52a1\u7684\u51c6\u786e\u6027\u3002<\/li>\n<\/ul>\n<p><strong>2\u3001\u6df1\u5ea6\u5b66\u4e60\u5174\u8d77\uff082010\u5e74\u4ee3\u4e2d\u671f\uff09<\/strong><\/p>\n<ul>\n<li>\u968f\u7740\u6df1\u5ea6\u5b66\u4e60\u6280\u672f\u7684\u5feb\u901f\u53d1\u5c55\uff0c\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\uff08CNN\uff09\u548c\u5faa\u73af\u795e\u7ecf\u7f51\u7edc\uff08RNN\uff09\u88ab\u5e7f\u6cdb\u5e94\u7528\u4e8e\u56fe\u50cf\u548c\u6587\u672c\u5904\u7406\u3002<\/li>\n<li>\u7814\u7a76\u8005\u63d0\u51fa\u4e86\u591a\u6a21\u6001\u5d4c\u5165\uff08embedding\uff09\u65b9\u6cd5\uff0c\u5c06\u4e0d\u540c\u6a21\u6001\u7684\u6570\u636e\u6620\u5c04\u5230\u540c\u4e00\u7a7a\u95f4\u4e2d\u3002<\/li>\n<\/ul>\n<p><strong>3\u3001Transformer\u67b6\u6784\u7684\u5f15\u5165\uff082017\u5e74\u53ca\u4ee5\u540e\uff09<\/strong><\/p>\n<ul>\n<li>Transformer\u67b6\u6784\u7684\u51fa\u73b0\u9769\u547d\u6027\u5730\u6539\u53d8\u4e86\u81ea\u7136\u8bed\u8a00\u5904\u7406\u548c\u8ba1\u7b97\u673a\u89c6\u89c9\u9886\u57df\u3002<\/li>\n<li>\u7ed3\u5408Transformer\u7684\u591a\u6a21\u6001\u6a21\u578b\uff08\u5982CLIP\u3001DALL-E\u7b49\uff09\u5f00\u59cb\u51fa\u73b0\uff0c\u80fd\u591f\u540c\u65f6\u5904\u7406\u56fe\u50cf\u548c\u6587\u672c\u3002<\/li>\n<\/ul>\n<p><strong>4\u3001\u5f53\u524d\u9636\u6bb5\uff082020\u5e74\u4ee3\uff09<\/strong><\/p>\n<ul>\n<li>\u591a\u6a21\u6001\u5927\u6a21\u578b\u5982GPT-4\u3001MUM\u3001Flamingo\u7b49\u76f8\u7ee7\u53d1\u5e03\uff0c\u8fd9\u4e9b\u6a21\u578b\u5728\u591a\u4e2a\u6a21\u6001\u4e0a\u8868\u73b0\u51fa\u8272\uff0c\u80fd\u591f\u8fdb\u884c\u590d\u6742\u7684\u4efb\u52a1\uff0c\u5982\u56fe\u50cf\u63cf\u8ff0\u3001\u6587\u672c\u751f\u6210\u3001\u95ee\u7b54\u7b49\u3002<\/li>\n<li>\u7814\u7a76\u8005\u4eec\u5173\u6ce8\u6a21\u578b\u7684\u53ef\u89e3\u91ca\u6027\u548c\u516c\u5e73\u6027\uff0c\u4ee5\u53ca\u5982\u4f55\u5728\u5b9e\u9645\u5e94\u7528\u4e2d\u66f4\u597d\u5730\u5229\u7528\u8fd9\u4e9b\u6a21\u578b\u3002<\/li>\n<\/ul>\n<h3><span class=\"ez-toc-section\" id=\"%E5%A4%9A%E6%A8%A1%E6%80%81%E5%A4%A7%E6%A8%A1%E5%9E%8B%E6%9E%B6%E6%9E%84%E7%BB%84%E6%88%90\"><\/span>\u591a\u6a21\u6001\u5927\u6a21\u578b\u67b6\u6784\u7ec4\u6210<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u591a\u6a21\u6001\u5927\u6a21\u578b\u67b6\u6784_\u6807\u6ce8.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u591a\u6a21\u6001\u5927\u6a21\u578b\u67b6\u6784_\u6807\u6ce8.png\" alt=\"\" \/><\/a><\/p>\n<h4><span class=\"ez-toc-section\" id=\"%E6%A8%A1%E6%80%81%E7%BC%96%E7%A0%81%E5%99%A8%E5%9B%BE%E2%91%A0\"><\/span>\u6a21\u6001\u7f16\u7801\u5668(\u56fe\u2460)<span class=\"ez-toc-section-end\"><\/span><\/h4>\n<ul>\n<li><strong>\u4f5c\u7528<\/strong>\uff1a\u662f\u5bf9\u6bcf\u79cd\u6a21\u6001(\u56fe\u7247\u3001\u97f3\u9891\u3001\u89c6\u9891)\u7684\u6570\u636e\u8fdb\u884c\u7279\u5f81\u63d0\u53d6\u548c\u7f16\u7801\uff0c\u5c06\u539f\u59cb\u8f93\u5165\u8f6c\u6362\u4e3a\u9ad8\u7ef4\u7279\u5f81\u8868\u793a\u3002<\/li>\n<li><strong>\u8bf4\u660e<\/strong>\uff1a\n<ul>\n<li>\u7279\u5f81\u63d0\u53d6\uff1a\u4e0d\u540c\u6a21\u6001\u7684\u6570\u636e\u4f7f\u7528\u4e0d\u540c\u7684\u7f51\u7edc\u67b6\u6784\u8fdb\u884c\u7279\u5f81\u63d0\u53d6\u3002\u4f8b\u5982\uff1a<\/li>\n<li>\u5d4c\u5165\u5c42\uff1a\u6bcf\u79cd\u6a21\u6001\u7684\u7279\u5f81\u7ecf\u8fc7\u5904\u7406\u540e\uff0c\u90fd\u4f1a\u88ab\u6620\u5c04\u5230\u4e00\u4e2a\u7edf\u4e00\u7684\u5d4c\u5165\u7a7a\u95f4\u4e2d\uff0c\u8fd9\u6837\u4e0d\u540c\u6a21\u6001\u7684\u7279\u5f81\u53ef\u4ee5\u5728\u540c\u4e00\u7a7a\u95f4\u5185\u8fdb\u884c\u6bd4\u8f83\u548c\u878d\u5408\u3002<\/li>\n<\/ul><\/li>\n<\/ul>\n<h4><span class=\"ez-toc-section\" id=\"%E6%A8%A1%E6%80%81%E8%BF%9E%E6%8E%A5%E5%99%A8%E5%9B%BE%E2%91%A1\"><\/span>\u6a21\u6001\u8fde\u63a5\u5668(\u56fe\u2461)<span class=\"ez-toc-section-end\"><\/span><\/h4>\n<ul>\n<li><strong>\u4f5c\u7528<\/strong>\uff1a\u5c06\u6a21\u6001\u7f16\u7801\u5668\u8f6c\u6362\u7684\u4e2d\u95f4\u8868\u8fbe\uff0c\u901a\u8fc7Connector\u6a21\u5757\uff0c\u5c06\u4e2d\u95f4\u8868\u8fbe\u8f6c\u6362\u4e3a\u4e0e\u5927\u8bed\u8a00\u6a21\u578b\u76f8\u540c\u7684\u8868\u8fbe<\/li>\n<li><strong>\u8bf4\u660e<\/strong>\uff1a\u6a21\u6001\u8fde\u63a5\u5668\u662f\u8bad\u7ec3\u5f62\u6210\u7684\uff0c\u5b83\u6709\u4e09\u79cd\u65b9\u5f0f\uff1a\n<h5><span class=\"ez-toc-section\" id=\"MLP_%E5%9F%BA%E4%BA%8E%E6%8A%95%E5%BD%B1%E7%9A%84%E8%BF%9E%E6%8E%A5%E5%99%A8%E5%9B%BE%E2%91%A4\"><\/span>MLP \u57fa\u4e8e\u6295\u5f71\u7684\u8fde\u63a5\u5668(\u56fe\u2464)<span class=\"ez-toc-section-end\"><\/span><\/h5>\n<p><strong>\u7b80\u8ff0<\/strong>\uff1a<code>\u591a\u5c42\u611f\u77e5\u673a<\/code> \u5c06\u7f16\u7801\u5668\u8f93\u51fa\u7684\u7279\u5f81\u6295\u5f71\u5230\u4e0eLLM\u7684\u8bcd\u5d4c\u5165\u76f8\u540c\u7684\u7ef4\u5ea6\u7a7a\u95f4\uff0c\u4f7f\u5f97\u7279\u5f81\u53ef\u4ee5\u76f4\u63a5\u4e0e\u6587\u672c\u4ee4\u724c\u4e00\u8d77\u88ab\u5339\u914d\u3002<\/p><\/li>\n<\/ul>\n<p><strong>\u539f\u7406<\/strong>\uff1a<\/p>\n<ul>\n<li><strong>\u8f93\u5165\u7279\u5f81<\/strong>\uff1a\u5c06\u6765\u81ea\u4e0d\u540c\u6a21\u6001\u7684\u7279\u5f81\u5411\u91cf\u62fc\u63a5\u5728\u4e00\u8d77\uff0c\u5f62\u6210\u4e00\u4e2a\u5927\u7684\u7279\u5f81\u5411\u91cf\u3002<\/li>\n<li><strong>\u5c42\u7ea7\u7ed3\u6784<\/strong>\uff1a\u901a\u8fc7\u591a\u4e2a\u5168\u8fde\u63a5\u5c42\uff08\u4e5f\u79f0\u4e3a\u7ebf\u6027\u5c42\uff09\u5bf9\u62fc\u63a5\u540e\u7684\u7279\u5f81\u8fdb\u884c\u5904\u7406\u3002\u6bcf\u4e2a\u5168\u8fde\u63a5\u5c42\u540e\u901a\u5e38\u4f1a\u52a0\u4e0a\u6fc0\u6d3b\u51fd\u6570\uff08\u5982ReLU\uff09\u4ee5\u5f15\u5165\u975e\u7ebf\u6027\u3002<\/li>\n<li><strong>\u8f93\u51fa<\/strong>\uff1a\u7ecf\u8fc7\u591a\u5c42\u5904\u7406\u540e\uff0c\u6700\u7ec8\u8f93\u51fa\u4e00\u4e2a\u878d\u5408\u7279\u5f81\u5411\u91cf\uff0c\u53ef\u4ee5\u7528\u4e8e\u540e\u7eed\u7684\u4efb\u52a1\uff08\u5982\u5206\u7c7b\u6216\u751f\u6210\uff09\u3002<\/li>\n<\/ul>\n<h5><span class=\"ez-toc-section\" id=\"Q-Former_%E5%9F%BA%E4%BA%8E%E6%9F%A5%E8%AF%A2%E7%9A%84%E8%BF%9E%E6%8E%A5%E5%99%A8%E5%9B%BE%E2%91%A5\"><\/span>Q-Former \u57fa\u4e8e\u67e5\u8be2\u7684\u8fde\u63a5\u5668(\u56fe\u2465)<span class=\"ez-toc-section-end\"><\/span><\/h5>\n<p><strong>\u7b80\u8ff0<\/strong>\uff1a\u4f7f\u7528\u4e00\u7ec4\u53ef\u5b66\u4e60\u7684\u67e5\u8be2\u4ee4\u724c\u6765\u52a8\u6001\u5730\u4ece\u7f16\u7801\u5668\u8f93\u51fa\u7684\u7279\u5f81\u4e2d\u63d0\u53d6\u4fe1\u606f\u3002\n<strong>\u539f\u7406<\/strong>\uff1a<\/p>\n<ul>\n<li><strong>\u67e5\u8be2\u3001\u952e\u3001\u503c<\/strong>\uff1aQ-Former \u5229\u7528\u67e5\u8be2\uff08Query\uff09\u3001\u952e\uff08Key\uff09\u548c\u503c\uff08Value\uff09\u6765\u8fdb\u884c\u7279\u5f81\u878d\u5408\u3002\u6bcf\u79cd\u6a21\u6001\u7684\u7279\u5f81\u88ab\u6620\u5c04\u4e3a\u67e5\u8be2\u3001\u952e\u548c\u503c\u3002<\/li>\n<li><strong>\u81ea\u6ce8\u610f\u529b\u673a\u5236<\/strong>\uff1a\u901a\u8fc7\u81ea\u6ce8\u610f\u529b\u673a\u5236\uff0cQ-Former \u53ef\u4ee5\u52a8\u6001\u5730\u4e3a\u4e0d\u540c\u6a21\u6001\u7684\u7279\u5f81\u5206\u914d\u4e0d\u540c\u7684\u6743\u91cd\uff0c\u4ece\u800c\u66f4\u597d\u5730\u6355\u6349\u6a21\u6001\u4e4b\u95f4\u7684\u5173\u7cfb\u3002<\/li>\n<li><strong>\u878d\u5408\u8fc7\u7a0b<\/strong>\uff1a\u5bf9\u4e8e\u6bcf\u4e00\u4e2a\u67e5\u8be2\uff0c\u8ba1\u7b97\u5176\u4e0e\u6240\u6709\u952e\u7684\u76f8\u4f3c\u5ea6\uff0c\u5e76\u6839\u636e\u76f8\u4f3c\u5ea6\u52a0\u6743\u6c42\u548c\u5bf9\u5e94\u7684\u503c\uff0c\u751f\u6210\u878d\u5408\u540e\u7684\u7279\u5f81\u8868\u793a\u3002<\/li>\n<\/ul>\n<h5><span class=\"ez-toc-section\" id=\"MH-Attn_%E5%9F%BA%E4%BA%8E%E8%9E%8D%E5%90%88%E7%9A%84%E8%BF%9E%E6%8E%A5%E5%99%A8%E5%9B%BE%E2%91%A6\"><\/span>MH-Attn \u57fa\u4e8e\u878d\u5408\u7684\u8fde\u63a5\u5668(\u56fe\u2466)<span class=\"ez-toc-section-end\"><\/span><\/h5>\n<p><strong>\u7b80\u8ff0<\/strong>\uff1a\u5728LLM\u5185\u90e8\u5b9e\u73b0\u7279\u5f81\u7ea7\u522b\u878d\u5408\uff0c\u5141\u8bb8\u6587\u672c\u7279\u5f81\u548c\u89c6\u89c9\u7279\u5f81\u5728\u6a21\u578b\u5185\u90e8\u8fdb\u884c\u66f4\u6df1\u5165\u7684\u4ea4\u4e92\u548c\u6574\u5408\u3002<\/p>\n<h4><span class=\"ez-toc-section\" id=\"%E6%A8%A1%E6%80%81%E7%94%9F%E6%88%90%E5%99%A8%E5%9B%BE%E2%91%A3\"><\/span>\u6a21\u6001\u751f\u6210\u5668(\u56fe\u2463)<span class=\"ez-toc-section-end\"><\/span><\/h4>\n<ul>\n<li>\u53ef\u9009\u7ec4\u4ef6\uff0c\u5b83\u53ef\u4ee5\u9644\u52a0\u5230LLM\u4e0a\uff0c\u7528\u4e8e\u751f\u6210\u9664\u6587\u672c\u4e4b\u5916\u7684\u5176\u4ed6\u6a21\u6001\uff0c\u5982\uff1a\u56fe\u7247\u3001\u97f3\u9891\u3001\u89c6\u9891\u7b49<\/li>\n<\/ul>\n<h2><span class=\"ez-toc-section\" id=\"%E5%A4%9A%E6%A8%A1%E6%80%81%E5%A4%A7%E6%A8%A1%E5%9E%8B%E8%AE%AD%E7%BB%83\"><\/span>\u591a\u6a21\u6001\u5927\u6a21\u578b\u8bad\u7ec3<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u4e0e\u5927\u8bed\u8a00\u6a21\u578b\u7684\u8bad\u7ec3\u8fc7\u7a0b\u7c7b\u4f3c<a href=\"https:\/\/17aitech.com\/?p=13611\">\u300a\u3010\u8bfe\u7a0b\u603b\u7ed3\u3011day24\uff08\u4e0a\uff09\uff1a\u5927\u6a21\u578b\u4e09\u9636\u6bb5\u8bad\u7ec3\u65b9\u6cd5(LLaMa Factory)\u300b<\/a>\uff0c\u591a\u6a21\u6001\u5927\u6a21\u578b\u4e5f\u6709 <code>\u9884\u8bad\u7ec3(Pre-train)<\/code>\u3001<code>\u5fae\u8c03(fine-tune)<\/code>\u3001<code>\u504f\u597d\u5bf9\u9f50(RLHF)<\/code> \u4e09\u4e2a\u8fc7\u7a0b\u3002<\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E7%AC%AC%E4%B8%80%E9%98%B6%E6%AE%B5%EF%BC%9A%E9%A2%84%E8%AE%AD%E7%BB%83Pre-train\"><\/span>\u7b2c\u4e00\u9636\u6bb5\uff1a\u9884\u8bad\u7ec3(Pre-train)<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><strong>\u9884\u8bad\u7ec3\u76ee\u7684<\/strong>\uff1a<\/p>\n<ul>\n<li>\u5bf9\u9f50\u6a21\u6001<\/li>\n<li>\u63d0\u4f9b\u4e16\u754c\u77e5\u8bc6<\/li>\n<\/ul>\n<p><strong>\u9884\u8bad\u7ec3\u6a21\u677f<\/strong>\uff1a<\/p>\n<pre><code class=\"language-bash\">Input: &lt;image&gt;\nResponse: {caption}<\/code><\/pre>\n<p><strong>\u9884\u8bad\u7ec3\u6570\u636e\u96c6<\/strong>\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u9884\u8bad\u7ec3\u6570\u636e\u96c6\u8868\u683c.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u9884\u8bad\u7ec3\u6570\u636e\u96c6\u8868\u683c.png\" alt=\"\" \/><\/a><\/p>\n<p>\u901a\u8fc7\u4e0a\u56fe\u53ef\u4ee5\u770b\u5230\uff0c\u9884\u8bad\u7ec3\u7684\u6570\u636e\u96c6\u7c7b\u578b\u633a\u591a\u7684\uff0c\u6709 \u7c97\u7c92\u5ea6\u7684\u56fe\u50cf-\u6587\u5b57\u6570\u636e(coarse-grained image-text)\u3001\u7ec6\u7c92\u5ea6\u7684\u56fe\u50cf-\u6587\u5b57\u6570\u636e(fine-grained image-text)\u3001\u89c6\u9891-\u6587\u5b57\u6570\u636e(video-text)\u3001\u97f3\u9891-\u6587\u5b57\u6570\u636e(audio-text)\u7b49\u3002<\/p>\n<p>\u4e3a\u4e86\u5bf9\u4ee5\u4e0a\u6570\u636e\u96c6\u6709\u4e2a\u76f4\u89c2\u4e86\u89e3\uff0c\u6211\u4eec\u6311\u51e0\u79cd\u6570\u636e\u96c6\u67e5\u770b\uff1a<\/p>\n<h4><span class=\"ez-toc-section\" id=\"%E7%B2%97%E7%B2%92%E5%BA%A6%E5%9B%BE%E5%83%8F-%E6%96%87%E5%AD%97%E6%95%B0%E6%8D%AE%E9%9B%86%E4%B8%80%E8%A7%88\"><\/span>\u7c97\u7c92\u5ea6\u56fe\u50cf-\u6587\u5b57\u6570\u636e\u96c6\u4e00\u89c8<span class=\"ez-toc-section-end\"><\/span><\/h4>\n<p><strong>\u6570\u636e\u96c6\u540d\u79f0<\/strong>\uff1aCC3M\n<strong>\u6570\u636e\u96c6\u5730\u5740<\/strong>\uff1a<a href=\"https:\/\/huggingface.co\/datasets\/pixparse\/cc3m-wds\">https:\/\/huggingface.co\/datasets\/pixparse\/cc3m-wds<\/a>\n<strong>\u6570\u636e\u96c6\u8bf4\u660e<\/strong>\uff1a\u8be5\u6570\u636e\u96c6\u4e3b\u8981\u662f\u7531\u56fe\u7247\u548c\u5bf9\u56fe\u7247\u7684\u63cf\u8ff0\u7ec4\u6210\u3002\n<strong>\u6570\u636e\u96c6\u622a\u56fe<\/strong>\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/CC3M\u6570\u636e\u96c6.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/CC3M\u6570\u636e\u96c6.png\" alt=\"\" \/><\/a><\/p>\n<h4><span class=\"ez-toc-section\" id=\"%E7%BB%86%E7%B2%92%E5%BA%A6%E5%9B%BE%E5%83%8F-%E6%96%87%E5%AD%97%E6%95%B0%E6%8D%AE%E9%9B%86%E4%B8%80%E8%A7%88\"><\/span>\u7ec6\u7c92\u5ea6\u56fe\u50cf-\u6587\u5b57\u6570\u636e\u96c6\u4e00\u89c8<span class=\"ez-toc-section-end\"><\/span><\/h4>\n<p><strong>\u6570\u636e\u96c6\u540d\u79f0<\/strong>\uff1aSHAREgpt4v-pt\n<strong>\u6570\u636e\u96c6\u5730\u5740<\/strong>\uff1a<a href=\"https:\/\/huggingface.co\/datasets\/Lin-Chen\/ShareGPT4V\/viewer\/ShareGPT4V\">https:\/\/huggingface.co\/datasets\/Lin-Chen\/ShareGPT4V\/viewer\/ShareGPT4V<\/a>\n<strong>\u6570\u636e\u96c6\u8bf4\u660e<\/strong>\uff1aSHAREgpt4v-pt \u6570\u636e\u96c6\u662f\u4e00\u4e2a\u4e13\u95e8\u7528\u4e8e\u591a\u6a21\u6001\u5927\u6a21\u578b\u8bad\u7ec3\u548c\u8bc4\u4f30\u7684\u6570\u636e\u96c6\uff0c\u5176\u6570\u636e\u96c6\u6709\u975e\u5e38\u8be6\u7ec6\u7684\u63cf\u8ff0\uff0c\u4ee5\u4e0b\u662fSHAREgpt4v-pt\u4e0eCOCO\u63cf\u8ff0\u7684\u5bf9\u6bd4\u3002\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/SHAREgpt4\u63cf\u8ff0\u5bf9\u6bd4.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/SHAREgpt4\u63cf\u8ff0\u5bf9\u6bd4.png\" alt=\"\" \/><\/a>\n<strong>\u6570\u636e\u96c6\u622a\u56fe<\/strong>\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/SHAREgpt4v.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/SHAREgpt4v.png\" alt=\"\" \/><\/a><\/p>\n<h4><span class=\"ez-toc-section\" id=\"%E8%A7%86%E9%A2%91-%E6%96%87%E5%AD%97%E6%95%B0%E6%8D%AE%E9%9B%86%E4%B8%80%E8%A7%88\"><\/span>\u89c6\u9891-\u6587\u5b57\u6570\u636e\u96c6\u4e00\u89c8<span class=\"ez-toc-section-end\"><\/span><\/h4>\n<p><strong>\u6570\u636e\u96c6\u540d\u79f0<\/strong>\uff1aMSRVTT\n<strong>\u6570\u636e\u96c6\u5730\u5740<\/strong>\uff1a<a href=\"https:\/\/huggingface.co\/datasets\/AlexZigma\/msr-vtt\">https:\/\/huggingface.co\/datasets\/AlexZigma\/msr-vtt<\/a>\n<strong>\u6570\u636e\u96c6\u8bf4\u660e<\/strong>\uff1aMSRVTT \u662f\u4e00\u4e2a\u89c6\u9891-\u6587\u672c\u6570\u636e\u96c6\uff0c\u5b83\u4e3b\u8981\u5185\u5bb9\u6709\u89c6\u9891\u5730\u5740\u3001\u89c6\u9891\u8d77\u59cb\u65f6\u95f4\u3001\u7ed3\u675f\u65f6\u95f4\u4ee5\u53ca\u89c6\u9891\u5e27\u6587\u5b57\u63cf\u8ff0\u6784\u6210\u3002\n<strong>\u6570\u636e\u96c6\u622a\u56fe<\/strong>\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/MSRVTT\u6570\u636e\u96c6.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/MSRVTT\u6570\u636e\u96c6.png\" alt=\"\" \/><\/a><\/p>\n<p>\u5176\u4e2dvideo0\u7684\u89c6\u9891\u5177\u4f53\u5185\u5bb9\u4e3a:\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/video0\u7684\u622a\u56fe.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/video0\u7684\u622a\u56fe.png\" alt=\"\" \/><\/a><\/p>\n<h4><span class=\"ez-toc-section\" id=\"%E9%9F%B3%E9%A2%91-%E6%96%87%E5%AD%97%E6%95%B0%E6%8D%AE%E9%9B%86%E4%B8%80%E8%A7%88\"><\/span>\u97f3\u9891-\u6587\u5b57\u6570\u636e\u96c6\u4e00\u89c8<span class=\"ez-toc-section-end\"><\/span><\/h4>\n<p><strong>\u6570\u636e\u96c6\u540d\u79f0<\/strong>\uff1awavCaps\n<strong>\u6570\u636e\u96c6\u5730\u5740<\/strong>\uff1a<a href=\"https:\/\/huggingface.co\/datasets\/cvssp\/WavCaps?row=0\">https:\/\/huggingface.co\/datasets\/cvssp\/WavCaps?row=0<\/a>\n<strong>\u6570\u636e\u96c6\u8bf4\u660e<\/strong>\uff1awavCaps \u662f\u4e00\u4e2a\u97f3\u9891-\u6587\u672c\u6570\u636e\u96c6\uff0c\u5b83\u4e3b\u8981\u5305\u542bFreeSound \u6570\u636e\u5e93\u5177\u4f53\u97f3\u9891\u7247\u6bb5\u3001\u5bf9\u5e94\u7684\u63cf\u8ff0\u6587\u5b57\u4fe1\u606f\u7f16\u53f7\u3002<\/p>\n<blockquote>\n<p>FreeSound \u662f\u4e00\u4e2a\u5f00\u653e\u7684\u97f3\u9891\u5171\u4eab\u5e73\u53f0\uff0c\u7528\u6237\u53ef\u4ee5\u4e0a\u4f20\u548c\u4e0b\u8f7d\u5404\u79cd\u7c7b\u578b\u7684\u97f3\u9891\u6587\u4ef6\uff0c\u6db5\u76d6\u81ea\u7136\u58f0\u97f3\u3001\u73af\u5883\u97f3\u6548\u3001\u97f3\u4e50\u7247\u6bb5\u7b49\u3002<\/p>\n<\/blockquote>\n<p><strong>\u6570\u636e\u96c6\u622a\u56fe<\/strong>\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/wavCaps\u6570\u636e\u96c6.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/wavCaps\u6570\u636e\u96c6.png\" alt=\"\" \/><\/a><\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E7%AC%AC%E4%BA%8C%E9%98%B6%E6%AE%B5%EF%BC%9A%E5%BE%AE%E8%B0%83fine-tune\"><\/span>\u7b2c\u4e8c\u9636\u6bb5\uff1a\u5fae\u8c03(fine-tune)<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><strong>\u5fae\u8c03\u76ee\u7684<\/strong>\uff1a<\/p>\n<ul>\n<li>\u8ba9\u6a21\u578b\u66f4\u597d\u5730\u7406\u89e3\u7528\u6237\u7684\u6307\u4ee4\u5e76\u5b8c\u6210\u6240\u9700\u7684\u4efb\u52a1<\/li>\n<li>\u6cdb\u5316\u80fd\u529b\uff0c\u5c11(\u96f6)\u6837\u672c\u63a8\u7406<\/li>\n<\/ul>\n<p><strong>\u5fae\u8c03\u6a21\u677f<\/strong>\uff1a<\/p>\n<pre><code class=\"language-bash\">Instruction: &lt;instruction&gt;\nInput: {&lt;image&gt;, &lt;text&gt;}\nResponse: &lt;output&gt;<\/code><\/pre>\n<p><strong>\u5fae\u8c03\u6570\u636e\u96c6<\/strong>\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u5fae\u8c03\u6570\u636e\u96c6\u8868\u683c.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u5fae\u8c03\u6570\u636e\u96c6\u8868\u683c.png\" alt=\"\" \/><\/a><\/p>\n<p>\u5bf9\u4e0a\u8ff0\u5fae\u8c03\u6570\u636e\u96c6\uff0c\u6211\u4eec\u6311\u9009\u4e24\u4e2a\u5178\u578b\u7684\u8fdb\u884c\u4e86\u89e3\u3002<\/p>\n<h4><span class=\"ez-toc-section\" id=\"LLaVa-Instruct\"><\/span>LLaVa-Instruct<span class=\"ez-toc-section-end\"><\/span><\/h4>\n<p><strong>\u6570\u636e\u96c6\u540d\u79f0<\/strong>\uff1aLLaVa-Instruct\n<strong>\u6570\u636e\u96c6\u5730\u5740<\/strong>\uff1a<a href=\"https:\/\/huggingface.co\/datasets\/liuhaotian\/LLaVA-Instruct-150K\">https:\/\/huggingface.co\/datasets\/liuhaotian\/LLaVA-Instruct-150K<\/a>\n<strong>\u6570\u636e\u96c6\u8bf4\u660e<\/strong>\uff1aLLaVa-Instruct \u6570\u636e\u96c6\u662f\u4e00\u4e2a\u7528\u4e8e\u8bad\u7ec3\u548c\u8bc4\u4f30\u591a\u6a21\u6001\u6a21\u578b\u7684\u6570\u636e\u96c6\uff0c\u5c24\u5176\u5173\u6ce8\u4e8e\u89c6\u89c9\u548c\u8bed\u8a00\u4e4b\u95f4\u7684\u4ea4\u4e92\u3002\u5b83\u65e8\u5728\u901a\u8fc7\u63d0\u4f9b\u6307\u4ee4\u548c\u76f8\u5e94\u7684\u89c6\u89c9\u5185\u5bb9\uff0c\u5e2e\u52a9\u6a21\u578b\u7406\u89e3\u548c\u751f\u6210\u4e0e\u89c6\u89c9\u4fe1\u606f\u76f8\u5173\u7684\u6587\u672c\u3002\n<strong>\u6570\u636e\u96c6\u622a\u56fe<\/strong>\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/LLaVa\u6570\u636e\u96c6\u622a\u56fe.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/LLaVa\u6570\u636e\u96c6\u622a\u56fe.png\" alt=\"\" \/><\/a><\/p>\n<h4><span class=\"ez-toc-section\" id=\"Video-ChatGPT\"><\/span>Video-ChatGPT<span class=\"ez-toc-section-end\"><\/span><\/h4>\n<p><strong>\u6570\u636e\u96c6\u540d\u79f0<\/strong>\uff1aVideo-ChatGPT\n<strong>\u6570\u636e\u96c6\u5730\u5740<\/strong>\uff1a<a href=\"https:\/\/huggingface.co\/datasets\/MBZUAI\/VideoInstruct-100K\">https:\/\/huggingface.co\/datasets\/MBZUAI\/VideoInstruct-100K<\/a>\n<strong>\u6570\u636e\u96c6\u8bf4\u660e<\/strong>\uff1aVideo-ChatGPT \u6570\u636e\u96c6\u662f\u4e00\u4e2a\u4e13\u95e8\u7528\u4e8e\u89c6\u9891\u7406\u89e3\u548c\u5bf9\u8bdd\u751f\u6210\u7684\u591a\u6a21\u6001\u6570\u636e\u96c6\uff0c\u5176\u6570\u636e\u96c6\u4e3b\u8981\u662f\u7531\u89c6\u9891id\u3001\u89c6\u9891\u7684\u95ee\u9898\u548c\u89c6\u9891\u95ee\u9898\u56de\u7b54\u4e09\u4e2a\u90e8\u5206\u7ec4\u6210\u3002\n<strong>\u6570\u636e\u96c6\u622a\u56fe<\/strong>\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/Video-ChatGPT\u6570\u636e\u96c6\u622a\u56fe.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/Video-ChatGPT\u6570\u636e\u96c6\u622a\u56fe.png\" alt=\"\" \/><\/a><\/p>\n<h4><span class=\"ez-toc-section\" id=\"Clotho\"><\/span>Clotho<span class=\"ez-toc-section-end\"><\/span><\/h4>\n<p><strong>\u6570\u636e\u96c6\u540d\u79f0<\/strong>\uff1aClotho\n<strong>\u6570\u636e\u96c6\u5730\u5740<\/strong>\uff1a<a href=\"https:\/\/paperswithcode.com\/dataset\/clotho\">https:\/\/paperswithcode.com\/dataset\/clotho<\/a>\n<strong>\u6570\u636e\u96c6\u8bf4\u660e<\/strong>\uff1a\u4e00\u4e2a\u7528\u4e8e\u97f3\u9891\u7406\u89e3\u548c\u751f\u6210\u7684\u591a\u6a21\u6001\u6570\u636e\u96c6\uff0c\u7279\u522b\u5173\u6ce8\u4e8e\u97f3\u9891\u63cf\u8ff0\u7684\u751f\u6210\u548c\u97f3\u9891\u5185\u5bb9\u7684\u7406\u89e3\u3002\n<strong>\u6570\u636e\u96c6\u622a\u56fe<\/strong>\uff1a\n(\u7531\u4e8e\u7f51\u7edc\u4e0d\u7a33\u5b9a\uff0c\u6682\u672a\u627e\u5230\u5bf9\u5e94\u6570\u636e\u96c6\u7684\u5177\u4f53\u5185\u5bb9)<\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E7%AC%AC%E4%B8%89%E9%98%B6%E6%AE%B5%EF%BC%9A%E5%81%8F%E5%A5%BD%E5%AF%B9%E9%BD%90\"><\/span>\u7b2c\u4e09\u9636\u6bb5\uff1a\u504f\u597d\u5bf9\u9f50<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><strong>\u504f\u597d\u5bf9\u9f50\u76ee\u7684<\/strong>\uff1a<\/p>\n<ul>\n<li>\u5bf9\u9f50\u5fae\u8c03\u6570\u636e<\/li>\n<\/ul>\n<p><strong>\u504f\u597d\u5bf9\u9f50\u6570\u636e\u96c6<\/strong>\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u504f\u597d\u5bf9\u9f50\u6570\u636e\u96c6.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u504f\u597d\u5bf9\u9f50\u6570\u636e\u96c6.png\" alt=\"\" \/><\/a><\/p>\n\n<!--more-->\n\n<h2><span class=\"ez-toc-section\" id=\"Qwen2%E7%9A%84%E9%83%A8%E7%BD%B2%E5%AE%9E%E8%B7%B5\"><\/span>Qwen2\u7684\u90e8\u7f72\u5b9e\u8df5<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<h3><span class=\"ez-toc-section\" id=\"%E5%90%AF%E5%8A%A8%E7%8E%AF%E5%A2%83\"><\/span>\u542f\u52a8\u73af\u5883<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u7b2c\u4e00\u6b65\uff1a\u542f\u52a8ModelScope\u5e73\u53f0\u7684PAI-DSW\u7684GPU\u73af\u5883<\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E6%A8%A1%E5%9E%8B%E4%B8%8B%E8%BD%BD\"><\/span>\u6a21\u578b\u4e0b\u8f7d<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u7b2c\u4e8c\u6b65\uff1a\u4e0b\u8f7d\u901a\u4e49\u5343\u95ee2-VL-2B-Instruct\u6a21\u578b<\/p>\n<pre><code class=\"language-bash\"># \u786e\u4fdd git lfs \u5df2\u5b89\u88c5\ngit lfs install\n\n# \u4e0b\u8f7d\u6a21\u578b\ngit clone https:\/\/www.modelscope.cn\/Qwen\/Qwen2-VL-2B-Instruct.git<\/code><\/pre>\n<h3><span class=\"ez-toc-section\" id=\"%E5%AE%89%E8%A3%85%E4%BE%9D%E8%B5%96\"><\/span>\u5b89\u88c5\u4f9d\u8d56<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<pre><code class=\"language-bash\">pip install qwen-vl-utils<\/code><\/pre>\n<h3><span class=\"ez-toc-section\" id=\"%E6%A8%A1%E5%9E%8B%E5%8A%A0%E8%BD%BD\"><\/span>\u6a21\u578b\u52a0\u8f7d<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<pre><code class=\"language-python\"># \u5f15\u5165\u76f8\u5173\u5e93\nfrom transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor\nfrom qwen_vl_utils import process_vision_info\n\n# \u8bbe\u7f6e\u6a21\u578b\u8def\u5f84\nmodel_dir = &quot;Qwen2-VL-2B-Instruct&quot;\n# \u52a0\u8f7d\u6a21\u578b\nmodel = Qwen2VLForConditionalGeneration.from_pretrained(\n    model_dir, torch_dtype=&quot;auto&quot;, device_map=&quot;auto&quot;\n)<\/code><\/pre>\n<p>\u8fd0\u884c\u7ed3\u679c\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/Qwen2_VL_\u52a0\u8f7d.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/Qwen2_VL_\u52a0\u8f7d.png\" alt=\"\" \/><\/a><\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E6%9E%84%E5%BB%BAprompt\"><\/span>\u6784\u5efaprompt<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<pre><code class=\"language-python\"># default processer\nprocessor = AutoProcessor.from_pretrained(model_dir)\n\nmessages = [\n    {\n        &quot;role&quot;: &quot;user&quot;,\n        &quot;content&quot;: [\n            {\n                &quot;type&quot;: &quot;image&quot;,\n                &quot;image&quot;: &quot;https:\/\/qianwen-res.oss-cn-beijing.aliyuncs.com\/Qwen-VL\/assets\/demo.jpeg&quot;,\n            },\n            {&quot;type&quot;: &quot;text&quot;, &quot;text&quot;: &quot;Describe this image.&quot;},\n        ],\n    }\n]<\/code><\/pre>\n<p>\u4e0a\u8ff0\u793a\u4f8b\u4ee3\u7801\u4e2d\uff0c\u63d0\u4f9b\u4e86\u4e00\u4e2ademo\u56fe\u7247\uff0c\u5176\u5185\u5bb9\u5982\u4e0b\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/demo.jpeg\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/demo.jpeg\" alt=\"\" \/><\/a><\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E5%87%86%E5%A4%87%E6%8E%A8%E7%90%86%E6%95%B0%E6%8D%AE\"><\/span>\u51c6\u5907\u63a8\u7406\u6570\u636e<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<pre><code class=\"language-python\"># \u51c6\u5907\u63a8\u7406\u6570\u636e\ntext = processor.apply_chat_template(\n    messages, tokenize=False, add_generation_prompt=True\n)\nimage_inputs, video_inputs = process_vision_info(messages)\ninputs = processor(\n    text=[text],\n    images=image_inputs,\n    videos=video_inputs,\n    padding=True,\n    return_tensors=&quot;pt&quot;,\n)\ninputs = inputs.to(&quot;cuda&quot;)<\/code><\/pre>\n<h3><span class=\"ez-toc-section\" id=\"%E6%A8%A1%E5%9E%8B%E6%8E%A8%E7%90%86\"><\/span>\u6a21\u578b\u63a8\u7406<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<pre><code class=\"language-python\"># \u6a21\u578b\u63a8\u7406\ngenerated_ids = model.generate(**inputs, max_new_tokens=128)\ngenerated_ids_trimmed = [\n    out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)\n]\noutput_text = processor.batch_decode(\n    generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False\n)\nprint(output_text)\n<\/code><\/pre>\n<p>\u8fd0\u884c\u7ed3\u679c\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/Qwen2_vl\u63a8\u7406\u7ed3\u679c.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/Qwen2_vl\u63a8\u7406\u7ed3\u679c.png\" alt=\"\" \/><\/a><\/p>\n<p>\u67e5\u770bGPU\u5360\u7528\u60c5\u51b5\uff1a\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/GPU\u5360\u7528\u60c5\u51b5.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/GPU\u5360\u7528\u60c5\u51b5.png\" alt=\"\" \/><\/a><\/p>\n<p>\u53ef\u4ee5\u770b\u52302B\u6a21\u578b\u7684GPU\u5360\u7528\u7387\u572870%\u5de6\u53f3\uff0c\u8fd8\u662f\u6bd4\u8f83\u5360\u663e\u5b58\u7684\u3002<\/p>\n<blockquote>\n<ul>\n<li>\u5982\u679cGPU\u663e\u5b58\u4e0d\u8db3\uff0c\u53ef\u4ee5\u4f7f\u7528\u91cf\u5316\u7248\u672c\uff0c\u5982Qwen-VL-Chat-Int4\u3002<\/li>\n<\/ul>\n<\/blockquote>\n<h2><span class=\"ez-toc-section\" id=\"%E9%87%8F%E5%8C%96%E5%86%85%E5%AE%B9%E7%AE%80%E4%BB%8B\"><\/span>\u91cf\u5316\u5185\u5bb9\u7b80\u4ecb<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<h3><span class=\"ez-toc-section\" id=\"%E9%87%8F%E5%8C%96%E7%9A%84%E5%AE%9A%E4%B9%89\"><\/span>\u91cf\u5316\u7684\u5b9a\u4e49<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><strong>\u91cf\u5316<\/strong>\u662f\u5c06\u9ad8\u7cbe\u5ea6\u6d6e\u70b9\u6570\u8868\u793a\uff08\u598232\u4f4d\u6d6e\u70b9\u6570\uff09\u8f6c\u6362\u4e3a\u4f4e\u7cbe\u5ea6\u8868\u793a\uff08\u598216\u4f4d\u30018\u4f4d\u6216\u66f4\u4f4e\u4f4d\u6570\uff09\u7684\u8fc7\u7a0b\u3002\u8fd9\u79cd\u8f6c\u6362\u53ef\u4ee5\u663e\u8457\u51cf\u5c11\u6a21\u578b\u7684\u5185\u5b58\u5360\u7528\u548c\u8ba1\u7b97\u590d\u6742\u5ea6\u3002<\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E9%87%8F%E5%8C%96%E7%9A%84%E8%83%8C%E6%99%AF\"><\/span>\u91cf\u5316\u7684\u80cc\u666f<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u7531\u4e8e\u6a21\u578b\u5728\u8fdb\u884cEmbedding\u5f52\u4e00\u5316\u65f6\uff0c\u4f1a\u5c06\u6570\u636e\u6620\u5c04\u5230[-1, 1]\u533a\u95f4\uff0c\u800c\u6b64\u65f6\u4f7f\u752832\u4f4d\u6d6e\u70b9\u6570\u6765\u8868\u793a\u6570\u636e\u65f6\uff0c\u4fdd\u5b58\u7684\u6570\u636e\u5730\u5740\u4f1a\u5b58\u5728\u4e00\u5b9a\u7a0b\u5ea6\u7684\u201c\u6d6a\u8d39\u201d\u3002\n\u56e0\u6b64\uff0c\u4e3a\u4e86\u51cf\u5c11\u6d6e\u70b9\u6570\u7684\u5b58\u50a8\u548c\u8ba1\u7b97\u5f00\u9500\uff0c\u5f15\u5165\u4e86\u91cf\u5316\u6280\u672f\uff0c\u5c06\u6d6e\u70b9\u6570\u8868\u793a\u8f6c\u6362\u4e3a\u6574\u6570\u8868\u793a\uff0c\u4ece\u800c\u8282\u7701\u4e86\u5b58\u50a8\u7a7a\u95f4\u548c\u8ba1\u7b97\u8d44\u6e90\u3002<\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E9%87%8F%E5%8C%96%E7%9A%84%E5%8E%9F%E7%90%86\"><\/span>\u91cf\u5316\u7684\u539f\u7406<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<ul>\n<li>\u6570\u503c\u8303\u56f4\u6620\u5c04\uff1a\u91cf\u5316\u8fc7\u7a0b\u4e2d\u9700\u8981\u786e\u5b9a\u6d6e\u70b9\u6570\u503c\u7684\u8303\u56f4\uff0c\u5e76\u5c06\u5176\u6620\u5c04\u5230\u4f4e\u7cbe\u5ea6\u8868\u793a\u7684\u53ef\u8868\u793a\u8303\u56f4\u5185\u3002\u4f8b\u5982\uff0c\u5c06\u4e00\u4e2a\u6d6e\u70b9\u6570\u503c\u8303\u56f4 [min, max] \u6620\u5c04\u5230 [0, 255]\uff088\u4f4d\u6574\u6570\uff09\u3002<\/li>\n<\/ul>\n<h3><span class=\"ez-toc-section\" id=\"%E9%87%8F%E5%8C%96%E7%9A%84%E4%BC%98%E7%82%B9\"><\/span>\u91cf\u5316\u7684\u4f18\u70b9<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<ul>\n<li>\u5b58\u50a8\u6548\u7387\uff1a\u51cf\u5c11\u6a21\u578b\u5927\u5c0f\uff0c\u964d\u4f4e\u5b58\u50a8\u9700\u6c42\uff0c\u4fbf\u4e8e\u90e8\u7f72\u5728\u8d44\u6e90\u6709\u9650\u7684\u8bbe\u5907\u4e0a\u3002<\/li>\n<li>\u8ba1\u7b97\u6548\u7387\uff1a\u4f4e\u7cbe\u5ea6\u8fd0\u7b97\u901a\u5e38\u6bd4\u9ad8\u7cbe\u5ea6\u8fd0\u7b97\u66f4\u5feb\uff0c\u80fd\u591f\u52a0\u901f\u63a8\u7406\u8fc7\u7a0b\uff0c\u63d0\u5347\u5b9e\u65f6\u6027\u3002<\/li>\n<li>\u80fd\u8017\u964d\u4f4e\uff1a\u4f4e\u7cbe\u5ea6\u8ba1\u7b97\u901a\u5e38\u6d88\u8017\u66f4\u5c11\u7684\u80fd\u91cf\uff0c\u6709\u52a9\u4e8e\u5728\u79fb\u52a8\u8bbe\u5907\u548c\u8fb9\u7f18\u8ba1\u7b97\u4e2d\u63d0\u5347\u80fd\u6548\u3002\n<h3><span class=\"ez-toc-section\" id=\"%E9%87%8F%E5%8C%96%E7%9A%84%E6%8C%91%E6%88%98\"><\/span>\u91cf\u5316\u7684\u6311\u6218<span class=\"ez-toc-section-end\"><\/span><\/h3><\/li>\n<li>\u6a21\u578b\u6027\u80fd\u4e0b\u964d\uff1a\u91cf\u5316\u53ef\u80fd\u5bfc\u81f4\u6a21\u578b\u7cbe\u5ea6\u4e0b\u964d\uff0c\u5c24\u5176\u662f\u5728\u5bf9\u7cbe\u5ea6\u8981\u6c42\u8f83\u9ad8\u7684\u4efb\u52a1\u4e2d\u3002<\/li>\n<li>\u91cf\u5316\u65b9\u6848\u7684\u9009\u62e9\uff1a\u4e0d\u540c\u7684\u91cf\u5316\u7b56\u7565\u5bf9\u6a21\u578b\u6027\u80fd\u7684\u5f71\u54cd\u4e0d\u540c\uff0c\u9700\u8981\u8fdb\u884c\u5b9e\u9a8c\u548c\u8c03\u4f18\u4ee5\u627e\u5230\u6700\u4f73\u65b9\u6848\u3002<\/li>\n<li>\u786c\u4ef6\u652f\u6301\uff1a\u5e76\u975e\u6240\u6709\u786c\u4ef6\u90fd\u652f\u6301\u4f4e\u7cbe\u5ea6\u8fd0\u7b97\uff0c\u53ef\u80fd\u9700\u8981\u7279\u5b9a\u7684\u786c\u4ef6\u6216\u8f6f\u4ef6\u4f18\u5316\u3002<\/li>\n<\/ul>\n<h2><span class=\"ez-toc-section\" id=\"%E5%86%85%E5%AE%B9%E5%B0%8F%E7%BB%93\"><\/span>\u5185\u5bb9\u5c0f\u7ed3<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<ul>\n<li>\u591a\u6a21\u6001\u5927\u6a21\u578b\uff08MM-LLM\uff09\u662f\u4e00\u79cd\u591a\u6a21\u6001\uff08\u6587\u672c\u3001\u56fe\u50cf\u3001\u97f3\u9891\u7b49\uff09\u7684 Large Language Model\uff08LLM\uff09\uff0c\u80fd\u591f\u540c\u65f6\u5904\u7406\u6587\u672c\u3001\u56fe\u50cf\u3001\u97f3\u9891\u7b49\u4e0d\u540c\u6a21\u6001\u7684\u4fe1\u606f\uff0c\u4ee5\u5b9e\u73b0\u66f4\u9ad8\u6548\u3001\u66f4\u9ad8\u8d28\u91cf\u7684\u591a\u6a21\u6001\u5bf9\u8bdd\u3002<\/li>\n<li>\u591a\u6a21\u6001\u5927\u6a21\u578b\u7684\u67b6\u6784\u7ec4\u6210\u4e3b\u8981\u589e\u52a0\u4e86\u6a21\u6001\u8fde\u63a5\u5668Connector\uff0c\u7528\u4e8e\u5c06\u4e0d\u540c\u6a21\u6001\u7684\u4fe1\u606f\u8fdb\u884c\u8fde\u63a5\uff0c\u4ee5\u589e\u5f3a\u6a21\u578b\u7684\u591a\u6a21\u6001\u80fd\u529b\u3002<\/li>\n<li>\u591a\u6a21\u6001\u5927\u6a21\u578b\u7684\u8bad\u7ec3\u8fc7\u7a0b\u4e5f\u5305\u542b\u4e09\u4e2a\u9636\u6bb5\uff1a\u9884\u8bad\u7ec3\u3001\u5fae\u8c03\u548c\u504f\u597d\u5bf9\u9f50\u3002<\/li>\n<li>\u591a\u6a21\u6001\u5927\u6a21\u578b\u7684\u8bad\u7ec3\u6570\u636e\u96c6\u4e0e\u5927\u6a21\u578b\u7c7b\u4f3c\uff0c\u53ea\u4e0d\u8fc7\u5728\u591a\u6a21\u6001\u6570\u636e\u4e0a\u8fdb\u884c\u4e86\u6269\u5c55\u3002<\/li>\n<li>\u591a\u6a21\u6001\u5927\u6a21\u578b\u7684\u90e8\u7f72\u63a8\u7406\u6bd4\u8f83\u7b80\u5355\uff0c\u9700\u8981\u4e0b\u8f7d\u4ee3\u7801\u3001\u5b89\u88c5\u4f9d\u8d56\u3001\u52a0\u8f7d\u6a21\u578b\uff0c\u5e76\u6784\u5efa\u63a8\u7406\u6570\u636e\uff0c\u7136\u540e\u8fdb\u884c\u63a8\u7406\u3002<\/li>\n<li>\u5982\u679cGPU\u663e\u5b58\u4e0d\u8db3\uff0c\u53ef\u4ee5\u4f7f\u7528\u91cf\u5316\u7248\u672c\uff0c\u5982Qwen-VL-Chat-Int4\u3002<\/li>\n<\/ul>\n<h2><span class=\"ez-toc-section\" id=\"%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99\"><\/span>\u53c2\u8003\u8d44\u6599<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p><a href=\"https:\/\/arxiv.org\/pdf\/2306.13549\">A Survey on Multimodal Large Language Models<\/a><\/p>\n<p align=\"center\">\u6b22\u8fce\u5173\u6ce8\u516c\u4f17\u53f7\u4ee5\u83b7\u5f97\u6700\u65b0\u7684\u6587\u7ae0\u548c\u65b0\u95fb<\/p>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u626b\u7801_\u641c\u7d22\u8054\u5408\u4f20\u64ad\u6837\u5f0f-\u767d\u8272\u7248.bmp\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u626b\u7801_\u641c\u7d22\u8054\u5408\u4f20\u64ad\u6837\u5f0f-\u767d\u8272\u7248.bmp\" alt=\"\" \/><\/a><\/p>","protected":false},"excerpt":{"rendered":"<p>\u524d\u8a00 \u968f\u77402024\u5e745\u670814\u65e5GPT-4 [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":32958,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"aside","meta":{"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","theme-transparent-header-meta":"default","adv-header-id-meta":"","stick-header-meta":"default","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"set","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"categories":[28],"tags":[73,68],"class_list":["post-32899","post","type-post","status-publish","format-aside","has-post-thumbnail","hentry","category-blog","tag-73","tag-68","post_format-post-format-aside"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v26.4 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>\u3010\u8bfe\u7a0b\u603b\u7ed3\u3011day31\uff1a\u591a\u6a21\u6001\u5927\u6a21\u578b\u521d\u6b65\u4e86\u89e3 - \u4e00\u8d77AI\u6280\u672f<\/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:\/\/17aitech.com\/?p=32899\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/17aitech.com\/?p=32899\",\"url\":\"https:\/\/17aitech.com\/?p=32899\",\"name\":\"\u3010\u8bfe\u7a0b\u603b\u7ed3\u3011day31\uff1a\u591a\u6a21\u6001\u5927\u6a21\u578b\u521d\u6b65\u4e86\u89e3 - \u4e00\u8d77AI\u6280\u672f\",\"isPartOf\":{\"@id\":\"https:\/\/17aitech.com\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/17aitech.com\/?p=32899#primaryimage\"},\"image\":{\"@id\":\"https:\/\/17aitech.com\/?p=32899#primaryimage\"},\"thumbnailUrl\":\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u591a\u6a21\u6001\u5927\u6a21\u578b\u67b6\u6784_\u6807\u6ce8.png\",\"datePublished\":\"2024-09-29T16:12:06+00:00\",\"dateModified\":\"2024-11-30T08:37:36+00:00\",\"author\":{\"@id\":\"https:\/\/17aitech.com\/#\/schema\/person\/3d23bb6f7f115fcefc9ae7803a691739\"},\"breadcrumb\":{\"@id\":\"https:\/\/17aitech.com\/?p=32899#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/17aitech.com\/?p=32899\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/17aitech.com\/?p=32899#primaryimage\",\"url\":\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u591a\u6a21\u6001\u5927\u6a21\u578b\u67b6\u6784_\u6807\u6ce8.png\",\"contentUrl\":\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u591a\u6a21\u6001\u5927\u6a21\u578b\u67b6\u6784_\u6807\u6ce8.png\",\"width\":1372,\"height\":1130},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/17aitech.com\/?p=32899#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\/\/17aitech.com\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"\u3010\u8bfe\u7a0b\u603b\u7ed3\u3011day31\uff1a\u591a\u6a21\u6001\u5927\u6a21\u578b\u521d\u6b65\u4e86\u89e3\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/17aitech.com\/#website\",\"url\":\"https:\/\/17aitech.com\/\",\"name\":\"\u4e00\u8d77AI\u6280\u672f\",\"description\":\"\u8ba9AI\u77e5\u8bc6\u89e6\u624b\u53ef\u53ca\",\"alternateName\":\"\u4e00\u8d77AI\u6280\u672f\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/17aitech.com\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/17aitech.com\/#\/schema\/person\/3d23bb6f7f115fcefc9ae7803a691739\",\"name\":\"Dongming\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/17aitech.com\/#\/schema\/person\/image\/\",\"url\":\"\/\/17aitech.com\/wp-content\/uploads\/member\/avatars\/238a0b923820dcc5.1732798681.jpg\",\"contentUrl\":\"\/\/17aitech.com\/wp-content\/uploads\/member\/avatars\/238a0b923820dcc5.1732798681.jpg\",\"caption\":\"Dongming\"},\"description\":\"\u89c1\u5929\u5730\uff0c\u89c1\u4f17\u751f\uff0c\u89c1\u81ea\u5df1\u3002\",\"sameAs\":[\"http:\/\/17aitech.com\"],\"url\":\"https:\/\/17aitech.com\/?page_id=33738&user=1\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"\u3010\u8bfe\u7a0b\u603b\u7ed3\u3011day31\uff1a\u591a\u6a21\u6001\u5927\u6a21\u578b\u521d\u6b65\u4e86\u89e3 - \u4e00\u8d77AI\u6280\u672f","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:\/\/17aitech.com\/?p=32899","schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/17aitech.com\/?p=32899","url":"https:\/\/17aitech.com\/?p=32899","name":"\u3010\u8bfe\u7a0b\u603b\u7ed3\u3011day31\uff1a\u591a\u6a21\u6001\u5927\u6a21\u578b\u521d\u6b65\u4e86\u89e3 - \u4e00\u8d77AI\u6280\u672f","isPartOf":{"@id":"https:\/\/17aitech.com\/#website"},"primaryImageOfPage":{"@id":"https:\/\/17aitech.com\/?p=32899#primaryimage"},"image":{"@id":"https:\/\/17aitech.com\/?p=32899#primaryimage"},"thumbnailUrl":"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u591a\u6a21\u6001\u5927\u6a21\u578b\u67b6\u6784_\u6807\u6ce8.png","datePublished":"2024-09-29T16:12:06+00:00","dateModified":"2024-11-30T08:37:36+00:00","author":{"@id":"https:\/\/17aitech.com\/#\/schema\/person\/3d23bb6f7f115fcefc9ae7803a691739"},"breadcrumb":{"@id":"https:\/\/17aitech.com\/?p=32899#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/17aitech.com\/?p=32899"]}]},{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/17aitech.com\/?p=32899#primaryimage","url":"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u591a\u6a21\u6001\u5927\u6a21\u578b\u67b6\u6784_\u6807\u6ce8.png","contentUrl":"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u591a\u6a21\u6001\u5927\u6a21\u578b\u67b6\u6784_\u6807\u6ce8.png","width":1372,"height":1130},{"@type":"BreadcrumbList","@id":"https:\/\/17aitech.com\/?p=32899#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/17aitech.com\/"},{"@type":"ListItem","position":2,"name":"\u3010\u8bfe\u7a0b\u603b\u7ed3\u3011day31\uff1a\u591a\u6a21\u6001\u5927\u6a21\u578b\u521d\u6b65\u4e86\u89e3"}]},{"@type":"WebSite","@id":"https:\/\/17aitech.com\/#website","url":"https:\/\/17aitech.com\/","name":"\u4e00\u8d77AI\u6280\u672f","description":"\u8ba9AI\u77e5\u8bc6\u89e6\u624b\u53ef\u53ca","alternateName":"\u4e00\u8d77AI\u6280\u672f","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/17aitech.com\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"zh-Hans"},{"@type":"Person","@id":"https:\/\/17aitech.com\/#\/schema\/person\/3d23bb6f7f115fcefc9ae7803a691739","name":"Dongming","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/17aitech.com\/#\/schema\/person\/image\/","url":"\/\/17aitech.com\/wp-content\/uploads\/member\/avatars\/238a0b923820dcc5.1732798681.jpg","contentUrl":"\/\/17aitech.com\/wp-content\/uploads\/member\/avatars\/238a0b923820dcc5.1732798681.jpg","caption":"Dongming"},"description":"\u89c1\u5929\u5730\uff0c\u89c1\u4f17\u751f\uff0c\u89c1\u81ea\u5df1\u3002","sameAs":["http:\/\/17aitech.com"],"url":"https:\/\/17aitech.com\/?page_id=33738&user=1"}]}},"_links":{"self":[{"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/posts\/32899","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/17aitech.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=32899"}],"version-history":[{"count":11,"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/posts\/32899\/revisions"}],"predecessor-version":[{"id":34128,"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/posts\/32899\/revisions\/34128"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/media\/32958"}],"wp:attachment":[{"href":"https:\/\/17aitech.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=32899"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/17aitech.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=32899"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/17aitech.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=32899"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}