{"id":1447,"date":"2022-02-02T05:10:00","date_gmt":"2022-02-01T20:10:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=1447"},"modified":"2022-02-02T05:10:00","modified_gmt":"2022-02-01T20:10:00","slug":"megatron-turing-nlg-530b%e3%81%ae%e8%a8%80%e8%aa%9e%e3%83%a2%e3%83%87%e3%83%ab","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=1447","title":{"rendered":"Megatron-Turing NLG: 530B\u306e\u8a00\u8a9e\u30e2\u30c7\u30eb"},"content":{"rendered":"\n<ul class=\"wp-block-list\"><li><strong>Using DeepSpeed and Megatron to Train Megatron-Turing NLG 530B, A Large-Scale Generative Language Model\u00a0<\/strong>[35.8]<br>\u672c\u7a3f\u3067\u306f,\u6700\u5927\u306e\u30e2\u30ce\u30ea\u30b7\u30c3\u30af\u30c8\u30e9\u30f3\u30b9\u30d5\u30a9\u30fc\u30de\u30fc\u30d9\u30fc\u30b9\u8a00\u8a9e\u30e2\u30c7\u30ebMegatron-Turing NLG 530B(MT-NLG)\u306e\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u306b\u3064\u3044\u3066\u8ff0\u3079\u308b\u3002 MT-NLG\u306f,\u3044\u304f\u3064\u304b\u306eNLP\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u306b\u304a\u3044\u3066,\u30bc\u30ed,\u30ef\u30f3\u30b7\u30e7\u30c3\u30c8,\u5c11\u6570\u30b7\u30e7\u30c3\u30c8\u306e\u5b66\u7fd2\u7cbe\u5ea6\u304c\u5411\u4e0a\u3057,\u65b0\u305f\u306a\u6700\u5148\u7aef\u7d50\u679c\u304c\u5f97\u3089\u308c\u305f\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2201.11990v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2201.11990v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Fri, 28 Jan 2022 08:59:57 GMT)<ul><li><a href=\"https:\/\/devneko.jp\/wordpress\/?p=814\">M6-10T \/ Yuan 1.0 \/ Megatron-Turing NLG 530B \u5de8\u5927\u306a\u4e8b\u524d\u5b66\u7fd2\u30e2\u30c7\u30eb \u2013 arXiv\u6700\u65b0\u8ad6\u6587\u306e\u7d39\u4ecb (devneko.jp)<\/a>\u306e\u8ad6\u6587\u30015300\u5104\u30d1\u30e9\u30e1\u30fc\u30bf\u306e\u5de8\u5927\u8a00\u8a9e\u30e2\u30c7\u30eb\u3002\u69d8\u3005\u306a\u6761\u4ef6\u3067GPT-3\u306e\u6027\u80fd\u3092\u4e0a\u56de\u308b\u3002<\/li><\/ul><ul><li>2\u30da\u30fc\u30b8\u76ee\u306e\u300cFigure 1: Trend of sizes of state-of-the-art NLP models with time.\u300d\u304b\u3089\u3082\u9806\u8abf\uff08\uff1f\uff09\u306b\u30e2\u30c7\u30eb\u30b5\u30a4\u30ba\u304c\u5927\u304d\u304f\u306a\u3063\u3066\u3044\u308b\u3053\u3068\u304c\u5206\u304b\u308b\u3002\u5b66\u7fd2\u7528\u306e\u30b7\u30b9\u30c6\u30e0\u69cb\u6210\u306f560 \u30ce\u30fc\u30c9\u306eDGX A100\u30011\u30ce\u30fc\u30c9\u3042\u305f\u308a 8\u3064\u306e NVIDIA 80-GB A100 GPU\u3001\u30d4\u30fc\u30af\u6027\u80fd\u306f1.4 exaFLOP\/s \uff0816-bit precision\uff09\u3068\u3059\u3054\u3044\u3002<\/li><\/ul><\/li><\/ul>\n","protected":false},"excerpt":{"rendered":"<p>Using DeepSpeed and Megatron to Train Megatron-Turing NLG 530B, A Large-Scale Generative Language Model\u00a0[35.8] &hellip; <a href=\"https:\/\/devneko.jp\/wordpress\/?p=1447\" class=\"more-link\"><span class=\"screen-reader-text\">&#8220;Megatron-Turing NLG: 530B\u306e\u8a00\u8a9e\u30e2\u30c7\u30eb&#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":[170,567],"class_list":["post-1447","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-gigantic-language-model","tag-567"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/1447","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=1447"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/1447\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1447"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=1447"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=1447"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}