{"id":710,"date":"2021-09-21T05:15:00","date_gmt":"2021-09-20T20:15:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=710"},"modified":"2021-09-21T05:15:00","modified_gmt":"2021-09-20T20:15:00","slug":"primerprimitives-searched-transformer-%e8%a8%80%e8%aa%9e%e3%83%a2%e3%83%87%e3%83%ab%e3%81%ae%e3%81%9f%e3%82%81%e3%81%ae%e5%8a%b9%e7%8e%87%e7%9a%84%e3%81%aa%e3%82%a2%e3%83%bc%e3%82%ad%e3%83%86","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=710","title":{"rendered":"Primer(PRIMitives searched transformER): \u8a00\u8a9e\u30e2\u30c7\u30eb\u306e\u305f\u3081\u306e\u52b9\u7387\u7684\u306a\u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3"},"content":{"rendered":"\n<ul class=\"wp-block-list\"><li><strong>Primer: Searching for Efficient Transformers for Language Modeling\u00a0<\/strong>[79.3]<br>\u5927\u578b\u30c8\u30e9\u30f3\u30b9\u30d5\u30a9\u30fc\u30de\u30fc\u30e2\u30c7\u30eb\u306e\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u3068\u63a8\u8ad6\u30b3\u30b9\u30c8\u306f\u6025\u901f\u306b\u5897\u52a0\u3057\u3001\u9ad8\u4fa1\u306b\u306a\u3063\u3066\u3044\u308b\u3002 \u3053\u3053\u3067\u306f\u3001\u3088\u308a\u52b9\u7387\u7684\u306a\u5909\u7a2e\u3092\u63a2\u3059\u3053\u3068\u306b\u3088\u308a\u3001\u30c8\u30e9\u30f3\u30b9\u30d5\u30a9\u30fc\u30de\u30fc\u306e\u30b3\u30b9\u30c8\u3092\u524a\u6e1b\u3059\u308b\u3053\u3068\u3092\u76ee\u7684\u3068\u3059\u308b\u3002 \u79c1\u305f\u3061\u306f\u3001\u30aa\u30ea\u30b8\u30ca\u30eb\u306eTransformer\u3088\u308a\u3082\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u30b3\u30b9\u30c8\u306e\u5c11\u306a\u3044Primer\u3068\u3044\u3046\u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3\u3092\u7279\u5b9a\u3057\u307e\u3059\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2109.08668v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2109.08668v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Fri, 17 Sep 2021 17:50:39 GMT)<ul><li>Transformer\u306e\u69cb\u9020\u306b\u624b\u3092\u5165\u308c\u52b9\u7387\u306e\u9ad8\u3044\u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3\u3092\u9032\u5316\u7684\u306a\u624b\u6cd5\u3067\u63a2\u7d22\u3002500M\u30d1\u30e9\u30e1\u30fc\u30bf\u306eT5\u76f8\u5f53\u306e\u5b9f\u88c5\u30674.2\u500d\u306e\u5b66\u7fd2\u30b3\u30b9\u30c8\u524a\u6e1b\u3092\u9054\u6210\u3059\u308bPrimer\u3068\u3044\u3046\u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3\u3092\u898b\u3064\u3051\u305f\u3068\u306e\u5831\u544a\u3002\u4e3b\u8981\u306a\u5909\u66f4\u70b9\u306f\u300cSquared ReLU\u300d\u3068\u300cMDHA: Multi-DConv-Head Attention \uff08\u6df1\u3055\u65b9\u5411\u306econvolution layer\u3092Q, K, V projection\u306e\u5f8c\u306b\u914d\u7f6e\uff09\u300d\u3002Primer\u306f\u3053\u306e2\u70b9\u4ee5\u5916\u306b\u3082\u69d8\u3005\u306a\u4fee\u6b63\u304c\u52a0\u3048\u3089\u308c\u3066\u3044\u308b\u304c\u3001\u3053\u306e\u5909\u66f4\u306e\u307f\u3092\u884c\u3063\u305f\u30d0\u30fc\u30b8\u30e7\u30f3\uff08Primer-EZ\uff09\u3067\u3082\u9ad8\u3044\u52b9\u679c\u304c\u3042\u308b\u3068\u306e\u3053\u3068\u3002\u8457\u8005\u306f\u307e\u305aPrimer-EZ\u3067\u9ad8\u901f\u5316\u52b9\u679c\u3092\u8a66\u3057\u5fc5\u8981\u306b\u5fdc\u3058\u3066\u5b8c\u5168\u306aPrimer\u3092\u5c0e\u5165\u3059\u308b\u3053\u3068\u3092\u63a8\u5968\u3057\u3066\u3044\u308b\u3002<\/li><li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/google-research\/google-research\/tree\/master\/primer\">https:\/\/github.com\/google-research\/google-research\/tree\/master\/primer<\/a><\/li><\/ul><\/li><\/ul>\n","protected":false},"excerpt":{"rendered":"<p>Primer: Searching for Efficient Transformers for Language Modeling\u00a0[79.3]\u5927\u578b\u30c8\u30e9\u30f3\u30b9\u30d5\u30a9\u30fc\u30de\u30fc\u30e2\u30c7\u30eb\u306e\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u3068\u63a8\u8ad6\u30b3\u30b9\u30c8\u306f\u6025\u901f\u306b\u5897\u52a0\u3057\u3001\u9ad8\u4fa1 &hellip; <a href=\"https:\/\/devneko.jp\/wordpress\/?p=710\" class=\"more-link\"><span class=\"screen-reader-text\">&#8220;Primer(PRIMitives searched transformER): \u8a00\u8a9e\u30e2\u30c7\u30eb\u306e\u305f\u3081\u306e\u52b9\u7387\u7684\u306a\u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3&#8221; \u306e<\/span>\u7d9a\u304d\u3092\u8aad\u3080<\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[415,567],"class_list":["post-710","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-transformer","tag-567"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/710","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=710"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/710\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=710"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=710"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=710"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}