微软崔宏禹：GPT-4以后，不建议再用微调进行特定领域的模型适配

老虎证券2023-05-27

【微软崔宏禹：GPT-4以后，不建议再用微调进行特定领域的模型适配】5月26日，微软（中国）有限公司创新技术总经理崔宏禹在上海举行的向星力·未来数据技术峰会（FDTC）上回应了当下一个广泛被提及的问题，即如何用自己的数据训练自己的模型，将私域数据与GPT的能力相融合。崔宏禹解答道，“一般来说，我们推荐两种方式。一种是微调（在不改动预训练模型的基础上，在模型顶层增加分类器或一些特征映射，使微调后的...

网页链接

免责声明：本文观点仅代表作者个人观点，不构成本平台的投资建议，本平台不对文章信息准确性、完整性和及时性做出任何保证，亦不对因使用或信赖文章信息引发的任何损失承担责任。

精彩评论

我们需要你的真知灼见来填补这片空白

发表看法

{"i18n":{"language":"zh_CN"},"isChannel":false,"data":{"share":"https://www.laohu8.com/m/news/2338748205?lang=zh_CN&edition=full","thumbnail":"","is_english":false,"pubTime":"2023-05-27 15:07","share_image_url":"https://static.laohu8.com/9a95c1376e76363c1401fee7d3717173","id":"2338748205","market":"us","top_or_hot":-1,"title":"微软崔宏禹：GPT-4以后，不建议再用微调进行特定领域的模型适配","media":"老虎证券","content":"<div>\n<p>【微软崔宏禹：GPT-4以后，不建议再用微调进行特定领域的模型适配】5月26日，微软（中国）有限公司创新技术总经理崔宏禹在上海举行的向星力·未来数据技术峰会（FDTC）上回应了当下一个广泛被提及的问题，即如何用自己的数据训练自己的模型，将私域数据与GPT的能力相融合。崔宏禹解答道，“一般来说，我们推荐两种方式。一种是微调（在不改动预训练模型的基础上，在模型顶层增加分类器或一些特征映射，使微调后的...</p>\n\n<a href=\"http://gu.qq.com/resources/shy/news/detail-v2/index.html#/?id=nesSN20230527151728824aee5e&s=b\">网页链接</a>\n\n</div>\n","source":"tencent","html":"<!DOCTYPE html>\n<html>\n<head>\n<meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\" />\n<meta name=\"viewport\" content=\"width=device-width,initial-scale=1.0,minimum-scale=1.0,maximum-scale=1.0,user-scalable=no\"/>\n<meta name=\"format-detection\" content=\"telephone=no,email=no,address=no\" />\n<title>微软崔宏禹：GPT-4以后，不建议再用微调进行特定领域的模型适配</title>\n<style type=\"text/css\">\na,abbr,acronym,address,applet,article,aside,audio,b,big,blockquote,body,canvas,caption,center,cite,code,dd,del,details,dfn,div,dl,dt,\nem,embed,fieldset,figcaption,figure,footer,form,h1,h2,h3,h4,h5,h6,header,hgroup,html,i,iframe,img,ins,kbd,label,legend,li,mark,menu,nav,\nobject,ol,output,p,pre,q,ruby,s,samp,section,small,span,strike,strong,sub,summary,sup,table,tbody,td,tfoot,th,thead,time,tr,tt,u,ul,var,video{ font:inherit;margin:0;padding:0;vertical-align:baseline;border:0 }\nbody{ font-size:16px; line-height:1.5; color:#999; background:transparent; }\n.wrapper{ overflow:hidden;word-break:break-all;padding:10px; }\nh1,h2{ font-weight:normal; line-height:1.35; margin-bottom:.6em; }\nh3,h4,h5,h6{ line-height:1.35; margin-bottom:1em; }\nh1{ font-size:24px; }\nh2{ font-size:20px; }\nh3{ font-size:18px; }\nh4{ font-size:16px; }\nh5{ font-size:14px; }\nh6{ font-size:12px; }\np,ul,ol,blockquote,dl,table{ margin:1.2em 0; }\nul,ol{ margin-left:2em; }\nul{ list-style:disc; }\nol{ list-style:decimal; }\nli,li p{ margin:10px 0;}\nimg{ max-width:100%;display:block;margin:0 auto 1em; }\nblockquote{ color:#B5B2B1; border-left:3px solid #aaa; padding:1em; }\nstrong,b{font-weight:bold;}\nem,i{font-style:italic;}\ntable{ width:100%;border-collapse:collapse;border-spacing:1px;margin:1em 0;font-size:.9em; }\nth,td{ padding:5px;text-align:left;border:1px solid #aaa; }\nth{ font-weight:bold;background:#5d5d5d; }\n.symbol-link{font-weight:bold;}\n/* header{ border-bottom:1px solid #494756; } */\n.title{ margin:0 0 8px;line-height:1.3;color:#ddd; }\n.meta {color:#5e5c6d;font-size:13px;margin:0 0 .5em; }\na{text-decoration:none; color:#2a4b87;}\n.meta .head { display: inline-block; overflow: hidden}\n.head .h-thumb { width: 30px; height: 30px; margin: 0; padding: 0; border-radius: 50%; float: left;}\n.head .h-content { margin: 0; padding: 0 0 0 9px; float: left;}\n.head .h-name {font-size: 13px; color: #eee; margin: 0;}\n.head .h-time {font-size: 11px; color: #7E829C; margin: 0;line-height: 11px;}\n.small {font-size: 12.5px; display: inline-block; transform: scale(0.9); -webkit-transform: scale(0.9); transform-origin: left; -webkit-transform-origin: left;}\n.smaller {font-size: 12.5px; display: inline-block; transform: scale(0.8); -webkit-transform: scale(0.8); transform-origin: left; -webkit-transform-origin: left;}\n.bt-text {font-size: 12px;margin: 1.5em 0 0 0}\n.bt-text p {margin: 0}\n</style>\n</head>\n<body>\n<div class=\"wrapper\">\n<header>\n<h2 class=\"title\">\n微软崔宏禹：GPT-4以后，不建议再用微调进行特定领域的模型适配\n</h2>\n\n<h4 class=\"meta\">\n\n\n2023-05-27 15:07 北京时间&nbsp;&nbsp;&nbsp;<a href=http://gu.qq.com/resources/shy/news/detail-v2/index.html#/?id=nesSN20230527151728824aee5e&s=b><strong>老虎证券</strong></a>\n\n\n</h4>\n\n</header>\n<article>\n<div>\n<p>【微软崔宏禹：GPT-4以后，不建议再用微调进行特定领域的模型适配】5月26日，微软（中国）有限公司创新技术总经理崔宏禹在上海举行的向星力·未来数据技术峰会（FDTC）上回应了当下一个广泛被提及的问题，即如何用自己的数据训练自己的模型，将私域数据与GPT的能力相融合。崔宏禹解答道，“一般来说，我们推荐两种方式。一种是微调（在不改动预训练模型的基础上，在模型顶层增加分类器或一些特征映射，使微调后的...</p>\n\n<a href=\"http://gu.qq.com/resources/shy/news/detail-v2/index.html#/?id=nesSN20230527151728824aee5e&s=b\">网页链接</a>\n\n</div>\n\n\n</article>\n</div>\n</body>\n</html>\n","isBrief":false,"type":0,"news_type":1,"symbol":"BK4577","symbol_name":"网络游戏","start_time":0,"source_url":"http://gu.qq.com/resources/shy/news/detail-v2/index.html#/?id=nesSN20230527151728824aee5e&s=b","article_id":"2338748205","we_media_id":null,"thumbnails":[],"rights":{"source":"tencent","url":"http://gu.qq.com/resources/shy/news/detail-v2/index.html#/?id=nesSN20230527151728824aee5e&s=b","rn_cache_url":null,"customStyle":"body{padding-top:10px;}#news_title{font-weight:bold;#titleStyle#;}#news_description span{font-size:12px;#descriptionStyle#;}.footer-note{#statement#}","selectors":".mod-LoadTzbdNews, body","filters":".relate-stock, .hot-list, .recom-box, .wx-sou","directOrigin":true},"url":"https://stock-news.laohu8.com/highlight/detail?id=2338748205","pubTimestamp":1685171240,"sourceInfo":{"source_id":"tencent","name":"腾讯"},"weMediaInfo":null,"summary":"【微软崔宏禹：GPT-4以后，不建议再用微调进行特定领域的模型适配】5月26日，微软（中国）有限公司创新技术总经理崔宏禹在上海举行的向星力·未来数据技术峰会（FDTC）上回应了当下一个广泛被提及的问题，即如何用自己的数据训练自己的模型，将私域数据与GPT的能力相融合。崔宏禹解答道，“一般来说，我们推荐两种方式。一种是微调（在不改动预训练模型的基础上，在模型顶层增加分类器或一些特征映射，使微调后的模型能够更贴合实际使用场景），但在GPT-4以后，我们已经不建议或者已经不让大家使用微调的方式了。我们更推荐的是嵌入（将一个内容实体映射为低维向量，从而可以获得内容之间的相似度）的方式。”（澎湃）","collect":0,"end_time":0,"defaultTopTitle":"qq.com","property":[],"viewcount":null,"language":"zh","relate_stocks":{"BK4577":"网络游戏","LU0079474960.USD":"联博美国增长基金A","LU0072462426.USD":"贝莱德全球配置 A2","BK4538":"云计算","MSFT":"微软","BK4579":"人工智能","LU0080751232.USD":"富达环球多元动力基金A","BK4588":"碎股","IE00B1XK9C88.USD":"PINEBRIDGE US LARGE CAP RESEARCH ENHANCED \"A\" (USD) ACC","LU0056508442.USD":"贝莱德世界科技基金A2","IE00BSNM7G36.USD":"NEUBERGER BERMAN SYSTEMATIC GLOBAL SUSTAINABLE VALUE \"A\" (USD) ACC","LU0109392836.USD":"富兰克林科技股A","BK4097":"系统软件","BK4581":"高盛持仓","BK4504":"桥水持仓","IE0004445015.USD":"JANUS HENDERSON BALANCED \"A2\" (USD) ACC","LU0011850046.USD":"贝莱德全球长线股票 A2 USD","LU0061474705.USD":"THREADNEEDLE (LUX) GLOBAL DYNAMIC REAL RETURN \"AU\" (USD) ACC","IE00BJJMRX11.SGD":"Janus Henderson Balanced A Acc SGD","BK4548":"巴美列捷福持仓","LU0158827948.USD":"ALLIANZ GLOBAL SUSTAINABILITY \"A\" (USD) INC","IE00B1BXHZ80.USD":"Legg Mason ClearBridge - US Appreciation A Acc USD","LU0289960550.SGD":"AB FCP I - GLOBAL EQUITY BLEND PORTFOLIO 'A' (SGD) ACC","IE00BLSP4452.SGD":"Legg Mason ClearBridge - Tactical Dividend Income A Mdis SGD-H Plus","LU0082616367.USD":"摩根大通美国科技A（dist）","LU0198837287.USD":"UBS (LUX) EQUITY SICAV - USA GROWTH \"P\" (USD) ACC","LU0276348264.USD":"THREADNEEDLE (LUX) GLOBAL DYNAMIC REAL RETURN\"AUP\" (USD) INC","BK4516":"特朗普概念","LU0061475181.USD":"THREADNEEDLE (LUX) AMERICAN \"AU\" (USD) ACC","IE0034235188.USD":"PINEBRIDGE GLOBAL FOCUS EQUITY \"A\" (USD) ACC","BK4554":"元宇宙及AR概念","LU0127658192.USD":"EASTSPRING INVESTMENTS GLOBAL TECHNOLOGY \"A\" (USD) ACC","IE00BKVL7J92.USD":"Legg Mason ClearBridge - US Equity Sustainability Leaders A Acc USD","IE0009356076.USD":"JANUS HENDERSON GLOBAL TECHNOLOGY AND INNOVATION \"A2\" (USD) ACC","IE00BJTD4N35.SGD":"Neuberger Berman US Long Short Equity A1  Acc SGD-H","LU0289739343.SGD":"SUSTAINABLE GLOBAL THEMATIC PORTFOLIO \"A\" (SGD) ACC","BK4534":"瑞士信贷持仓","BK4585":"ETF&股票定投概念","IE00B7KXQ091.USD":"Janus Henderson Balanced A Inc USD","IE00BLSP4239.USD":"Legg Mason ClearBridge - Tactical Dividend Income A Mdis USD Plus","LU0234570918.USD":"高盛全球核心股票组合Acc Close","BK4533":"AQR资本管理(全球第二大对冲基金)","IE00B775SV38.USD":"NEUBERGER BERMAN US MULTICAP OPPORTUNITIES \"A\" (USD) ACC","IE00BFSS7M15.SGD":"Janus Henderson Balanced A Acc SGD-H","LU0109391861.USD":"富兰克林美国机遇基金A Acc","BK4587":"ChatGPT概念","IE00B3S45H60.SGD":"Neuberger Berman US Multicap Opportunities A Acc SGD-H","LU0234572021.USD":"高盛美国核心股票组合Acc","IE0004445239.USD":"JANUS HENDERSON US FORTY \"A2\" (USD) ACC","BK4535":"淡马锡持仓"},"translate_title":"Microsoft Cui Hongyu: After GPT-4, it is not recommended to use fine-tuning for model adaptation in specific fields","themeId":null,"isJumpTheme":false,"ttsUrl":null,"symbols_score_info":{"MSFT":0.9},"content_text":"【微软崔宏禹：GPT-4以后，不建议再用微调进行特定领域的模型适配】5月26日，微软（中国）有限公司创新技术总经理崔宏禹在上海举行的向星力·未来数据技术峰会（FDTC）上回应了当下一个广泛被提及的问题，即如何用自己的数据训练自己的模型，将私域数据与GPT的能力相融合。崔宏禹解答道，“一般来说，我们推荐两种方式。一种是微调（在不改动预训练模型的基础上，在模型顶层增加分类器或一些特征映射，使微调后的模型能够更贴合实际使用场景），但在GPT-4以后，我们已经不建议或者已经不让大家使用微调的方式了。我们更推荐的是嵌入（将一个内容实体映射为低维向量，从而可以获得内容之间的相似度）的方式。”（澎湃）","kind":null,"is_publish_news":null,"is_publish_highlight":null,"is_publish_live":null,"is_publish_wemedia":null,"editions":null,"symbols":[],"gpt_button":0},"commentList":[],"isCommentEnd":true,"newsSizeData":{"likeSize":0,"commentSize":0,"repostSize":0,"favoriteSize":0,"likeStatus":false,"favoriteStatus":false},"APP":{"userAgent":"Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthropic.com)","isDev":false,"isTTM":false,"deviceId":"web-server-community-laohu8-v3","version":"4.22.4","shortVersion":"4.22.4","platform":"web","vendor":"web","appName":"laohu8","isIOS":false,"isAndroid":false,"isTiger":false,"isTHS":false,"isWeiXin":false,"isWeiXinMini":false,"isWeiBo":false,"isQQ":false,"isBaiduSwan":false,"isBaiduBox":false,"isDingTalk":false,"isToutiao":false,"isOnePlus":false,"isHuaWei":false,"isXiaomi":false,"isXiaomiWebView":false,"isOppo":false,"isVivo":false,"isSamsung":false,"isMobile":false},"href":"/m/news/2338748205"}