{"id":2751,"date":"2023-01-02T05:43:00","date_gmt":"2023-01-01T20:43:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=2751"},"modified":"2023-01-02T05:43:00","modified_gmt":"2023-01-01T20:43:00","slug":"cramming-training-a-language-model-on-a-single-gpu-in-one-day","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=2751","title":{"rendered":"Cramming: Training a Language Model on a Single GPU in One Day"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Cramming: Training a Language Model on a Single GPU in One Day\u00a0<\/strong>[64.2]<br>\u8a00\u8a9e\u30e2\u30c7\u30ea\u30f3\u30b0\u306e\u6700\u8fd1\u306e\u30c8\u30ec\u30f3\u30c9\u306f\u3001\u30b9\u30b1\u30fc\u30ea\u30f3\u30b0\u306b\u3088\u308b\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u5411\u4e0a\u306b\u7126\u70b9\u3092\u5f53\u3066\u3066\u3044\u308b\u3002 \u6211\u3005\u306f,1\u3064\u306e\u30b3\u30f3\u30b7\u30e5\u30fc\u30deGPU\u4e0a\u30671\u65e5\u9593,\u30de\u30b9\u30af\u4ed8\u304d\u8a00\u8a9e\u30e2\u30c7\u30eb\u3092\u7528\u3044\u3066\u30bc\u30ed\u304b\u3089\u5b8c\u5168\u306b\u8a13\u7df4\u3055\u308c\u305f\u30c8\u30e9\u30f3\u30b9\u30d5\u30a9\u30fc\u30de\u30fc\u30d9\u30fc\u30b9\u8a00\u8a9e\u30e2\u30c7\u30eb\u3067\u9054\u6210\u53ef\u80fd\u306a\u30c0\u30a6\u30f3\u30b9\u30c8\u30ea\u30fc\u30e0\u6027\u80fd\u306b\u3064\u3044\u3066\u691c\u8a0e\u3057\u305f\u3002 \u3053\u306e\u5236\u7d04\u3055\u308c\u305f\u8a2d\u5b9a\u3067\u3042\u3063\u3066\u3082\u3001\u5927\u898f\u6a21\u8a2d\u5b9a\u3067\u89b3\u6e2c\u3055\u308c\u308b\u30b9\u30b1\u30fc\u30ea\u30f3\u30b0\u306e\u6cd5\u5247\u306b\u5bc6\u63a5\u306b\u5f93\u3046\u6027\u80fd\u3092\u793a\u3059\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2212.14034v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2212.14034v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Wed, 28 Dec 2022 18:59:28 GMT)<\/li>\n\n\n\n<li>\u81ea\u7136\u8a00\u8a9e\u51e6\u7406\u306e\u30bf\u30b9\u30af\u306b\u3064\u3044\u3066\uff11GPU day\u3067\u3069\u3053\u307e\u3067\u6027\u80fd\u3092\u4f38\u3070\u305b\u308b\u304b\u3092\u691c\u8a3c\u3057\u305f\u8ad6\u6587\u3002\u975e\u5e38\u306b\u8208\u5473\u6df1\u3044\u8a2d\u5b9a\u3067\u5e83\u7bc4\u306a\u5b9f\u9a13\u304c\u3055\u308c\u3066\u3044\u308b\u3002<\/li>\n\n\n\n<li>\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u306b\u3088\u308b\u5dee\u30011 GPU day\u3068\u306f\u3044\u3048\u3001GPU\u306e\u7a2e\u985e\uff08\u2252\u8a08\u7b97\u8cc7\u6e90\uff09\u306b\u3088\u308b\u5dee\u306b\u3064\u3044\u3066\u3082\u9762\u767d\u3044\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/JonasGeiping\/cramming\">GitHub &#8211; JonasGeiping\/cramming: Cramming the training of a (BERT-type) language model into limited compute.<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[551,645],"class_list":["post-2751","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-551","tag-645"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/2751","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=2751"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/2751\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=2751"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=2751"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=2751"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}