{"id":6101,"date":"2025-01-27T06:00:00","date_gmt":"2025-01-26T21:00:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=6101"},"modified":"2025-01-27T06:00:00","modified_gmt":"2025-01-26T21:00:00","slug":"deepseek-r1","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=6101","title":{"rendered":"Deepseek R1\u3001Sky-T1\u3001TinyZero\u3001Kimi k1.5"},"content":{"rendered":"\n<p>\u5148\u9031\u3082\u5927\u304d\u306a\u30cb\u30e5\u30fc\u30b9\u304c\u591a\u304b\u3063\u305f\u3002\u7279\u306bDeepSeek R1\u306f\u975e\u5e38\u306b\u9ad8\u3044\u6027\u80fd\u306eLarge Reasoning Model\u3067\u3042\u308a\u3001\u3057\u304b\u3082\u3001\u30aa\u30fc\u30d7\u30f3\u306a\u30e2\u30c7\u30eb\u3067\u3042\u308b\u3053\u3068\u304c\u885d\u6483\u7684\u3060\u3063\u305f\u3002Deepseek R1 Zero\u306f\u5f37\u5316\u5b66\u7fd2\u306b\u3088\u3063\u3066\u6027\u80fd\u3092\u4e0a\u3052\u3066\u3044\u308b\u3053\u3068\u3082\u7279\u5fb4\u7684\u3067\u3042\u308b\u3002Kimi k1.5\u3082\u8fd1\u3044\u767a\u60f3\u3067\u69cb\u7bc9\u3055\u308c\u305f\u30e2\u30c7\u30eb\u3067\u5f37\u5316\u5b66\u7fd2\u306e\u6709\u52b9\u6027\u3092\u793a\u3057\u3066\u3044\u308b\u3088\u3046\u306b\u898b\u3048\u308b\u3002<\/p>\n\n\n\n<p>DeepSeek R1\u306e\u904e\u7a0b\u3067\u69cb\u7bc9\u3057\u305f\u30c7\u30fc\u30bf\u3092\u7528\u3044Qwen\u3084Llama\u3092\u5f37\u5316\u3057\u305f\u30e2\u30c7\u30eb\u3082\u5927\u304d\u304f\u6027\u80fd\u3092\u4e0a\u3052\u3066\u3044\u308b\u306e\u304c\u9a5a\u304d\u3002\u84b8\u7559\u304c\u8a31\u53ef\u3055\u308c\u3066\u3044\u308b\u30e9\u30a4\u30bb\u30f3\u30b9\u3067\u3042\u308a\u3001\u5408\u6210\u30c7\u30fc\u30bf\u3092\u69cb\u7bc9\u3059\u308b\u5143\u30e2\u30c7\u30eb\u3068\u3057\u3066\u3082\u6709\u529b\u305d\u3046\u3002<\/p>\n\n\n\n<p>o1\u30e9\u30a4\u30af\u306a\u30aa\u30fc\u30d7\u30f3\u30e2\u30c7\u30eb\u3068\u3057\u3066\u306f<a href=\"https:\/\/novasky-ai.github.io\/posts\/sky-t1\/\">Sky-T1: Train your own O1 preview model within $450<\/a>\u3084<a href=\"https:\/\/github.com\/Jiayi-Pan\/TinyZero\">GitHub &#8211; Jiayi-Pan\/TinyZero<\/a>\uff08<a href=\"https:\/\/x.com\/jiayi_pirate\/status\/1882839370505621655\">X\u30e6\u30fc\u30b6\u30fc\u306eJiayi Pan\u3055\u3093: \u300cWe reproduced DeepSeek R1-Zero in the CountDown game, and it just works Through RL, the 3B base LM develops self-verification and search abilities all on its own You can experience the Ahah moment yourself for &lt; $30 Code: https:\/\/t.co\/B2IsN1PrXV Here&#8217;s what we learned \ud83e\uddf5 https:\/\/t.co\/43BVYMmS8X\u300d \/ X<\/a>\uff09\u3082\u8208\u5473\u6df1\u3044\u3002<\/p>\n\n\n\n<p>\u305d\u308c\u4ee5\u5916\u306b\u3082OpenAI Operator\uff08<a href=\"https:\/\/openai.com\/index\/introducing-operator\/\">Introducing Operator research preview | OpenAI<\/a>\uff09\u306fGUI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u306e\u840c\u82bd\u3092\u611f\u3058\u3055\u305b\u308b\u3002<\/p>\n\n\n\n<p>\u30aa\u30fc\u30d7\u30f3\u30e2\u30c7\u30eb\u306e\u76db\u308a\u4e0a\u304c\u308a\u306e\u4e2d\u3001OpenAI\u304cLLM\u30b3\u30a2\u3060\u3051\u3067\u306f\u306a\u304f\u5468\u8fba\u9818\u57df\u306b\u624b\u3092\u51fa\u305d\u3046\u3068\u3057\u3066\u3044\u308b\u3088\u3046\u306b\u3082\u898b\u3048\u3066\u9762\u767d\u3044\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning&nbsp;<\/strong>[147.2]<br>\u7b2c\u4e00\u4e16\u4ee3\u306e\u63a8\u8ad6\u30e2\u30c7\u30eb\u3067\u3042\u308bDeepSeek-R1-Zero\u3068DeepSeek-R1\u3092\u7d39\u4ecb\u3057\u307e\u3059\u3002 DeepSeek-R1-Zero\u306f\u5927\u898f\u6a21\u306a\u5f37\u5316\u5b66\u7fd2\u3092\u901a\u3058\u3066\u8a13\u7df4\u3055\u308c\u3066\u3044\u308b\u3002 DeepSeek-R1\u306f\u3001RL\u306e\u524d\u306b\u30de\u30eb\u30c1\u30b9\u30c6\u30fc\u30b8\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u3068\u30b3\u30fc\u30eb\u30c9\u30b9\u30bf\u30fc\u30c8\u30c7\u30fc\u30bf\u3092\u7d44\u307f\u8fbc\u3093\u3067\u3044\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2501.12948v1\">\u8ad6\u6587<\/a>&nbsp;&nbsp;<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2501.12948v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>&nbsp; &nbsp;(Wed, 22 Jan 2025 15:19:35 GMT)<\/li>\n<\/ul>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Kimi k1.5: Scaling Reinforcement Learning with LLMs\u00a0<\/strong>[84.2]<br>\u6211\u3005\u306f\u3001\u5f37\u5316\u5b66\u7fd2\u3067\u8a13\u7df4\u3055\u308c\u305f\u6700\u65b0\u306e\u30de\u30eb\u30c1\u30e2\u30fc\u30c0\u30eb\u8a00\u8a9e\u30e2\u30c7\u30eb\u3001Kimi k1.5\u306e\u8a13\u7df4\u5b9f\u8df5\u306b\u3064\u3044\u3066\u5831\u544a\u3059\u308b\u3002 \u9577\u3044\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8\u30b9\u30b1\u30fc\u30ea\u30f3\u30b0\u3068\u6539\u5584\u3055\u308c\u305f\u30dd\u30ea\u30b7\u30fc\u6700\u9069\u5316\u624b\u6cd5\u304c\u3001\u6211\u3005\u306e\u30a2\u30d7\u30ed\u30fc\u30c1\u306e\u9375\u3068\u306a\u308b\u8981\u7d20\u3067\u3042\u308b\u3002 \u672c\u30b7\u30b9\u30c6\u30e0\u306f,\u8907\u6570\u306e\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3084\u30e2\u30c0\u30ea\u30c6\u30a3\u306b\u5bfe\u3057\u3066,\u6700\u5148\u7aef\u306e\u63a8\u8ad6\u6027\u80fd\u3092\u5b9f\u73fe\u3059\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2501.12599v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2501.12599v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Wed, 22 Jan 2025 02:48:14 GMT)<\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>\u5148\u9031\u3082\u5927\u304d\u306a\u30cb\u30e5\u30fc\u30b9\u304c\u591a\u304b\u3063\u305f\u3002\u7279\u306bDeepSeek R1\u306f\u975e\u5e38\u306b\u9ad8\u3044\u6027\u80fd\u306eLarge Reasoning Model\u3067\u3042\u308a\u3001\u3057\u304b\u3082\u3001\u30aa\u30fc\u30d7\u30f3\u306a\u30e2\u30c7\u30eb\u3067\u3042\u308b\u3053\u3068\u304c\u885d\u6483\u7684\u3060\u3063\u305f\u3002Deepseek R1 Zero\u306f\u5f37\u5316\u5b66\u7fd2\u306b &hellip; <a href=\"https:\/\/devneko.jp\/wordpress\/?p=6101\" class=\"more-link\"><span class=\"screen-reader-text\">&#8220;Deepseek R1\u3001Sky-T1\u3001TinyZero\u3001Kimi k1.5&#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":[232,283,576],"class_list":["post-6101","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-lrm","tag-o1","tag-576"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6101","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=6101"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6101\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=6101"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=6101"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=6101"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}