{"id":5164,"date":"2024-07-22T06:02:00","date_gmt":"2024-07-21T21:02:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=5164"},"modified":"2024-07-22T06:02:00","modified_gmt":"2024-07-21T21:02:00","slug":"gpt-4o-mini%e3%80%81mistralnemo%e3%80%81dclm-7b","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=5164","title":{"rendered":"GPT-4o mini\u3001MistralNeMo\u3001DCLM 7B\u3001Qwen2\u3001Granite"},"content":{"rendered":"\n<p>\u5148\u9031\u3082LLM\u95a2\u9023\u306e\u8a71\u984c\u306f\u591a\u304b\u3063\u305f\u3002<\/p>\n\n\n\n<p>GPT-3.5\u3088\u308a\u3082\u30b3\u30b9\u30c8\u30fb\u901f\u5ea6\u306b\u512a\u308c\u6027\u80fd\u3082\u9ad8\u3044GPT-4o mini\u306f\u30d3\u30b8\u30cd\u30b9\u3067\u306e\u6ce8\u76ee\u5ea6\u304c\u9ad8\u3044\u3002<\/p>\n\n\n\n<p>OSS\u95a2\u9023\u3067\u306e\u30a2\u30c3\u30d7\u30c7\u30fc\u30c8\u3082\u591a\u304b\u3063\u305f\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Mistral\u3068NVIDIA\u304c\u5354\u529b\u3057\u3066\u958b\u767a\u3057\u305f\u5c0f\u578b\u3067\u5f37\u529b\u306aMistral NeMo\uff08<a href=\"https:\/\/mistral.ai\/news\/mistral-nemo\/\">Mistral NeMo | Mistral AI | Frontier AI in your hands<\/a>\u3001<a href=\"https:\/\/huggingface.co\/mistralai\/Mistral-Nemo-Base-2407\">mistralai\/Mistral-Nemo-Base-2407 \u00b7 Hugging Face<\/a>\uff09<\/li>\n\n\n\n<li>Apple\u306b\u3088\u308bDCLM 7B\uff08<a href=\"https:\/\/huggingface.co\/apple\/DCLM-7B\">apple\/DCLM-7B \u00b7 Hugging Face<\/a>\uff06\u95a2\u9023\uff1a<a href=\"https:\/\/devneko.jp\/wordpress\/?p=5025\"><\/a><a href=\"https:\/\/devneko.jp\/wordpress\/?p=5025\">DataComp-LM: In search of the next generation of training sets for language models \u2013 arXiv\u6700\u65b0\u8ad6\u6587\u306e\u7d39\u4ecb (devneko.jp)<\/a>\uff09<\/li>\n\n\n\n<li>Qwen2\u306b\u3064\u3044\u3066\u306e\u30c6\u30af\u30cb\u30ab\u30eb\u30ec\u30dd\u30fc\u30c8\uff08Audio-Language\u542b\u3080\uff09<\/li>\n\n\n\n<li>\u9577\u3044\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8\u306b\u5bfe\u5fdc\u3057\u305fGranite<\/li>\n<\/ul>\n\n\n\n<p>\u4e0a\u8a18\u306b\u306f\u8981\u6ce8\u76ee\u3002\u516c\u958b\u30e2\u30c7\u30eb\u306e\u52d5\u304d\u3082\u901f\u3044\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Qwen2 Technical Report&nbsp;<\/strong>[139.8]<br>Qwen2\u306f\u3001\u524d\u6a5f\u7a2e\u306eQwen1.5\u3092\u542b\u3080\u3001\u3053\u308c\u307e\u3067\u306e\u307b\u3068\u3093\u3069\u306e\u30aa\u30fc\u30d7\u30f3\u30a6\u30a7\u30a4\u30c8\u30e2\u30c7\u30eb\u3092\u4e0a\u56de\u3063\u3066\u3044\u308b\u3002 \u8a00\u8a9e\u7406\u89e3\u3001\u751f\u6210\u3001\u591a\u8a00\u8a9e\u7fd2\u719f\u3001\u30b3\u30fc\u30c7\u30a3\u30f3\u30b0\u3001\u6570\u5b66\u3001\u63a8\u8ad6\u306b\u95a2\u3059\u308b\u69d8\u3005\u306a\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u306b\u304a\u3044\u3066\u3001\u30d7\u30ed\u30d7\u30e9\u30a4\u30a8\u30bf\u30ea\u306a\u30e2\u30c7\u30eb\u3068\u6bd4\u8f03\u3057\u3066\u3001\u7af6\u4e89\u529b\u306e\u3042\u308b\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u3092\u793a\u3059\u3002 Qwen2\u306f\u3001\u82f1\u8a9e\u3001\u4e2d\u56fd\u8a9e\u3001\u30b9\u30da\u30a4\u30f3\u8a9e\u3001\u30d5\u30e9\u30f3\u30b9\u8a9e\u3001\u30c9\u30a4\u30c4\u8a9e\u3001\u30a2\u30e9\u30d3\u30a2\u8a9e\u3001\u30ed\u30b7\u30a2\u8a9e\u3001\u97d3\u56fd\u8a9e\u3001\u65e5\u672c\u8a9e\u3001\u30bf\u30a4\u8a9e\u3001\u30d9\u30c8\u30ca\u30e0\u8a9e\u306a\u3069\u3001\u7d0430\u306e\u8a00\u8a9e\u3067\u719f\u7df4\u3057\u305f\u5805\u7262\u306a\u591a\u8a00\u8a9e\u6a5f\u80fd\u3092\u793a\u3057\u3066\u3044\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2407.10671v1\">\u8ad6\u6587<\/a>&nbsp;&nbsp;<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2407.10671v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>&nbsp; &nbsp;(Mon, 15 Jul 2024 12:35:42 GMT)<\/li>\n\n\n\n<li><a href=\"https:\/\/devneko.jp\/wordpress\/?p=4957\">GLM-4-9B, Qwen2 \u2013 arXiv\u6700\u65b0\u8ad6\u6587\u306e\u7d39\u4ecb (devneko.jp)<\/a>\u306e\u30c6\u30af\u30cb\u30ab\u30eb\u30ec\u30dd\u30fc\u30c8\u3002\u5f37\u529b\u306a\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u3092\u767a\u63ee\u3057\u3001\u591a\u8a00\u8a9e\u6027\u80fd\u3082\u9ad8\u3044\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/QwenLM\/Qwen2\">GitHub &#8211; QwenLM\/Qwen2: Qwen2 is the large language model series developed by Qwen team, Alibaba Cloud.<\/a><\/li>\n<\/ul>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Qwen2-Audio Technical Report&nbsp;<\/strong>[73.9]<br>\u672c\u7a3f\u3067\u306f,Qwen2-Audio\u3068\u547c\u3070\u308c\u308b\u5927\u898f\u6a21\u30aa\u30fc\u30c7\u30a3\u30aa\u8a00\u8a9e\u30e2\u30c7\u30eb\u3067\u3042\u308bQwen-Audio\u306e\u6700\u65b0\u52d5\u5411\u3092\u7d39\u4ecb\u3059\u308b\u3002 Qwen2-Audio\u306f\u3001\u69d8\u3005\u306a\u97f3\u58f0\u4fe1\u53f7\u5165\u529b\u3092\u53d7\u3051\u5165\u308c\u3001\u97f3\u58f0\u89e3\u6790\u3084\u97f3\u58f0\u6307\u793a\u306b\u5bfe\u3059\u308b\u76f4\u63a5\u30c6\u30ad\u30b9\u30c8\u5fdc\u7b54\u3092\u884c\u3046\u3053\u3068\u304c\u3067\u304d\u308b\u3002 \u6211\u3005\u306fQwen2-Audio\u306e\u6307\u793a\u8ffd\u5f93\u80fd\u529b\u3092\u9ad8\u3081\u3001\u97f3\u58f0\u30c1\u30e3\u30c3\u30c8\u3068\u97f3\u58f0\u5206\u6790\u306e\u305f\u3081\u306e2\u3064\u306e\u7570\u306a\u308b\u97f3\u58f0\u5bfe\u8a71\u30e2\u30fc\u30c9\u3092\u5b9f\u88c5\u3057\u305f\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2407.10759v1\">\u8ad6\u6587<\/a>&nbsp;&nbsp;<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2407.10759v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>&nbsp; &nbsp;(Mon, 15 Jul 2024 14:38:09 GMT)<\/li>\n\n\n\n<li>\u300cAccording to the evaluation results from AIR-Bench, Qwen2-Audio outperformed previous SOTAs, such as Gemini-1.5-pro, in tests focused on audio-centric instruction-following capabilities.\u300d\u3068\u5f37\u529b\u306a\u6027\u80fd\u3092\u4e3b\u5f35\u3059\u308bAudio-Language\u30e2\u30c7\u30eb\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/QwenLM\/Qwen2-Audio\">GitHub &#8211; QwenLM\/Qwen2-Audio: The official repo of Qwen2-Audio chat &amp; pretrained large audio language model proposed by Alibaba Cloud.<\/a><\/li>\n<\/ul>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Scaling Granite Code Models to 128K Context\u00a0<\/strong>[37.3]<br>\u672c\u7a3f\u3067\u306f,\u6700\u5927128K\u30c8\u30fc\u30af\u30f3\u306e\u52b9\u7387\u7684\u306a\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8\u30a6\u30a3\u30f3\u30c9\u30a6\u3092\u30b5\u30dd\u30fc\u30c8\u3059\u308b\u9577\u6587\u30b0\u30e9\u30ca\u30a4\u30c8\u7b26\u53f7\u30e2\u30c7\u30eb\u3092\u63d0\u6848\u3059\u308b\u3002 2K\/4K\u304b\u3089128K\u307e\u3067\u306eGranite 3B\/8B\u7b26\u53f7\u30e2\u30c7\u30eb\u306e\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8\u9577\u306e\u30b9\u30b1\u30fc\u30ea\u30f3\u30b0\u30bd\u30ea\u30e5\u30fc\u30b7\u30e7\u30f3\u306f\u3001\u8efd\u91cf\u306a\u7d99\u7d9a\u4e8b\u524d\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u3067\u69cb\u6210\u3055\u308c\u3066\u3044\u308b\u3002 \u79c1\u305f\u3061\u306f\u3001\u7814\u7a76\u3068\u5546\u7528\u306e\u4e21\u65b9\u306e\u305f\u3081\u306b\u3001Apache 2.0\u30e9\u30a4\u30bb\u30f3\u30b9\u306e\u4e0b\u3067\u3001\u9577\u3044\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8\u306eGranite\u30b3\u30fc\u30c9\u30e2\u30c7\u30eb\u3092\u30ea\u30ea\u30fc\u30b9\u3057\u307e\u3059\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2407.13739v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2407.13739v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Thu, 18 Jul 2024 17:46:02 GMT)<\/li>\n\n\n\n<li>IBM\u306eGranite\u3082128K\u3068\u9577\u3044\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8\u306b\u5bfe\u5fdc<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/ibm-granite\/granite-code-models\">GitHub &#8211; ibm-granite\/granite-code-models: Granite Code Models: A Family of Open Foundation Models for Code Intelligence<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>\u5148\u9031\u3082LLM\u95a2\u9023\u306e\u8a71\u984c\u306f\u591a\u304b\u3063\u305f\u3002 GPT-3.5\u3088\u308a\u3082\u30b3\u30b9\u30c8\u30fb\u901f\u5ea6\u306b\u512a\u308c\u6027\u80fd\u3082\u9ad8\u3044GPT-4o mini\u306f\u30d3\u30b8\u30cd\u30b9\u3067\u306e\u6ce8\u76ee\u5ea6\u304c\u9ad8\u3044\u3002 OSS\u95a2\u9023\u3067\u306e\u30a2\u30c3\u30d7\u30c7\u30fc\u30c8\u3082\u591a\u304b\u3063\u305f\u3002 \u4e0a\u8a18\u306b\u306f\u8981\u6ce8\u76ee\u3002\u516c\u958b\u30e2\u30c7\u30eb\u306e\u52d5\u304d\u3082\u901f\u3044\u3002<\/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":[223,293],"class_list":["post-5164","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-llm","tag-oss"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/5164","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=5164"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/5164\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=5164"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=5164"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=5164"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}