{"id":24049,"date":"2024-11-29T09:27:49","date_gmt":"2024-11-29T01:27:49","guid":{"rendered":"https:\/\/www.1ai.net\/?p=24049"},"modified":"2024-11-29T09:27:49","modified_gmt":"2024-11-29T01:27:49","slug":"%e4%b8%ad%e5%9b%bd%e7%9f%b3%e6%b2%b9%e5%8f%91%e5%b8%83-700-%e4%ba%bf%e5%8f%82%e6%95%b0%e6%98%86%e4%bb%91%e5%a4%a7%e6%a8%a1%e5%9e%8b%ef%bc%8c%e4%b8%8e%e4%b8%ad%e5%9b%bd%e7%a7%bb%e5%8a%a8%e3%80%81","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/24049.html","title":{"rendered":"PetroChina Releases 70 Billion Parameter Kunlun Large Model, Joins Forces with China Mobile, Huawei, and KDDI"},"content":{"rendered":"

according toChina National Petroleum Corporation<\/a>The report, dated November 28, is based on\"AI Pilot, Intelligent Energy.\"<\/strong>themed\u00a0<\/strong>70 billion parametersKunlun Large Model<\/a><\/strong>The conference on the construction results was held in Beijing. At the conference, PetroChina introduced the 70 billion parameter Kunlun model jointly built with China Mobile, Huawei and Tech Data, and displayed 43 innovative scenarios of specialized and general applications.<\/p>\n

\"0ddfd97fj00snoup300i2d000k800o3p\"<\/p>\n

In terms of the industry's big models, theQuorum Large Model Language Large Model Parameters Upgraded from 33 Billion to 70 Billion<\/strong>In addition, the professional capabilities of natural language processing, speech and visual recognition, multimodal interaction and scientific computing have been greatly improved; the training data of visual large model has been further enriched, and the newly added rock recognition function has effectively improved the level of intelligence in geological research; the newly trained multimodal large model with 16 billion parameters has provided rich text, graphic and audio multimodal interaction capabilities for the marketing and service business of \"People, Vehicles and Life\". The newly trained 16 billion parameter multimodal large model provides rich multimodal interaction capabilities of text, graphics and audio for marketing and service such as \"people, car and life\".<\/p>\n

In terms of specialized macromodels, theIt has constructed China's first large-scale model of exploration in all fields of specialization.<\/strong>The model covers 3 major models, namely seismic processing, seismic interpretation, and log processing interpretation, and it has greatly improved the generalization and accuracy compared with the traditional methods.<\/p>\n

In terms of scenario macromodeling, theOn-demand construction of 21 scenario models, each supporting multiple applications for one type of application scenario.<\/strong>. At the same time, Kunlun Big Model has established a scene dataset return mechanism to regularly converge the scene corpus into professional and industry datasets, gradually improving the professional and industry big model capabilities and reducing the number of scene big models.<\/p>\n

In addition, PetroChina has constructedOver 280TB of high quality data sets<\/strong>,Developed a centralized and unified AI platform for group companies.<\/strong>The unified management and co-scheduling capabilities of arithmetic, models, and applications are effectively strengthened.<\/p>\n

1AI has learned that Kunlun Big Model, the first big model in China's energy and chemical industry to pass the filing, officially started construction in May this year, with the parameter magnitude reaching 33 billion after three months.<\/p>","protected":false},"excerpt":{"rendered":"

According to PetroChina, on November 28, a conference on the results of the construction of the 70 billion-parameter Kunlun Large Model was held in Beijing with the theme of \"AI Pilotage\u30fbSmart Mapping of Energy\". At the conference, PetroChina introduced the 70 billion-parameter Kunlun Big Model jointly built with China Mobile, Huawei, and Tech Data Xunfei, and displayed 43 innovative scenarios of professional and general applications. In terms of industry big models, the parameters of Kunlun big model language big model have been upgraded from 33 billion to 70 billion, and the professional capabilities of natural language processing, speech and visual recognition, multimodal interaction, and scientific computation have been greatly improved; the training data of visual big model has been further enriched, and the newly added function of rock recognition has effectively improved the level of intelligence in geological research; the newly trained multimodal big model with 16 billion parameters has provided a new model for \"People and Vehicles<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[2836,2835],"collection":[],"class_list":["post-24049","post","type-post","status-publish","format-standard","hentry","category-news","tag-2836","tag-2835"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/24049","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=24049"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/24049\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=24049"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=24049"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=24049"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=24049"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}