{"id":6819,"date":"2024-04-01T09:28:44","date_gmt":"2024-04-01T01:28:44","guid":{"rendered":"https:\/\/www.1ai.net\/?p=6819"},"modified":"2024-04-01T09:28:44","modified_gmt":"2024-04-01T01:28:44","slug":"databricks-%e6%8e%a8%e5%87%ba-1320-%e4%ba%bf%e5%8f%82%e6%95%b0%e5%a4%a7%e8%af%ad%e8%a8%80%e6%a8%a1%e5%9e%8b-dbrx%ef%bc%8c%e5%8f%b7%e7%a7%b0%e7%8e%b0%e9%98%b6%e6%ae%b5%e6%9c%80%e5%bc%ba","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/6819.html","title":{"rendered":"Databricks launches DBRX, a 132 billion parameter large language model, known as &quot;the most powerful open source AI at this stage&quot;"},"content":{"rendered":"<p data-vmark=\"0e2e\"><a href=\"https:\/\/www.1ai.net\/en\/tag\/databricks\" title=\"[See articles with [Databricks] labels]\" target=\"_blank\" >Databricks<\/a> Recently, a universal<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%a4%a7%e8%af%ad%e8%a8%80%e6%a8%a1%e5%9e%8b\" title=\"[View articles tagged with [large language model]]\" target=\"_blank\" >Large Language Model<\/a> DBRX, known as &quot;the most powerful open source AI at present&quot;,<span class=\"accentTextColor\">It is said to surpass &quot;all other<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%bc%80%e6%ba%90%e6%a8%a1%e5%9e%8b\" title=\"[See articles with [open source model] labels]\" target=\"_blank\" >Open Source Model<\/a>\u201d<\/span>.<\/p>\n<p data-vmark=\"4790\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-6820\" title=\"818a17dd-8720-43d4-bf97-9b1824cc546a\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2024\/04\/818a17dd-8720-43d4-bf97-9b1824cc546a.png\" alt=\"818a17dd-8720-43d4-bf97-9b1824cc546a\" width=\"1440\" height=\"963\" \/><\/p>\n<p data-vmark=\"ad2e\">According to the official press release, DBRX is a large language model based on Transformer, using the MoE (Mixture of Experts) architecture.<span class=\"accentTextColor\">It has 132 billion parameters and is pre-trained on 12T Token source data.<\/span>.<\/p>\n<p data-vmark=\"e345\">The researchers tested this model and found that compared with the existing open source models such as LLaMA2-70B, Mixtral, and Grok-1 on the market, DBRX performed better in benchmarks such as language understanding (MMLU), programming (HumanEval), and mathematical logic (GSM8K). The official also mentioned that DBRX also surpassed OpenAI&#039;s GPT-3.5 in the above three benchmarks.<\/p>\n<p data-vmark=\"2cac\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-6821\" title=\"c285b5be-b7e7-409d-a7b2-ff81b9c89a2d\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2024\/04\/c285b5be-b7e7-409d-a7b2-ff81b9c89a2d.png\" alt=\"c285b5be-b7e7-409d-a7b2-ff81b9c89a2d\" width=\"1440\" height=\"857\" \/><\/p>\n<p data-vmark=\"1dc0\">Naveen Rao, vice president of Databricks AI, told TechCrunch that the company spent two months and $10 million to train DBRX. Although DBRX currently performs well in terms of overall results, the model currently requires four NVIDIA H100 GPUs to run, so there is still room for optimization.<\/p>","protected":false},"excerpt":{"rendered":"<p>Databricks recently launched DBRX, a general-purpose large language model, which is claimed to be \"the strongest open source AI available today\" and is said to outperform \"all open source models on the market\" in various benchmark tests. According to the official press release, DBRX is a Transformer-based large-language model that utilizes the MoE (Mixture of Experts) architecture, has 132 billion parameters, and is pre-trained on 12T Token of source data. The researchers tested the model and found that compared to open source models such as LLaMA2-70B, Mixtral, Grok-1, etc., which are already available in the market, DBRX is better in language understanding (MMLU), programming (MMLU), and language analysis (MMLU) than the other models on the market.<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[1906,706,862],"collection":[],"class_list":["post-6819","post","type-post","status-publish","format-standard","hentry","category-news","tag-databricks","tag-706","tag-862"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/6819","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=6819"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/6819\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=6819"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=6819"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=6819"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=6819"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}