{"id":27595,"date":"2025-01-22T21:50:41","date_gmt":"2025-01-22T13:50:41","guid":{"rendered":"https:\/\/www.1ai.net\/?p=27595"},"modified":"2025-01-22T21:50:41","modified_gmt":"2025-01-22T13:50:41","slug":"%e7%bd%91%e6%98%93%e6%9c%89%e9%81%93%e6%8e%a8%e5%87%ba%e5%9b%bd%e5%86%85%e9%a6%96%e4%b8%aa%e8%be%93%e5%87%ba%e5%88%86%e6%ad%a5%e5%bc%8f%e8%ae%b2%e8%a7%a3%e7%9a%84%e6%8e%a8%e7%90%86%e6%a8%a1%e5%9e%8b","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/27595.html","title":{"rendered":"NetEaseYouDao launches \"ZiYao-o1\", the first reasoning model that outputs step-by-step explanations in China"},"content":{"rendered":"<p>January 22nd.<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e7%bd%91%e6%98%93%e6%9c%89%e9%81%93\" title=\"[Sees articles with [accessible] labels]\" target=\"_blank\" >NetEase Youtube<\/a>today announced the launch of the country's first output step-by-step explanation of the<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e6%8e%a8%e7%90%86%e6%a8%a1%e5%9e%8b\" title=\"[View articles tagged with [inference model]]\" target=\"_blank\" >inference model<\/a>\"Tzu-o1\" and officially open-sourced.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-27596\" title=\"0fe3b0d3j00sqht3d0080d000ow00alp\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2025\/01\/0fe3b0d3j00sqht3d0080d000ow00alp.jpg\" alt=\"0fe3b0d3j00sqht3d0080d000ow00alp\" width=\"896\" height=\"381\" \/><\/p>\n<p>According to the official introduction, ZiYu-o1 is a 14B lightweight single model that supports deployment on consumer-grade graphics cards and uses thought chain technology to be able to provide a detailed problem solving process to achieve higher problem solving accuracy with strong logic and reasoning capabilities, as well as to provide logical reasoning in Chinese.<\/p>\n<p>NetEase said there are not many open source models available and the parameter scales are too large to run on consumer graphics cards with low video memory, even though low bit quantization technology is used to enable it to be deployed on a single card, but accordingly brings instability to the operation of the long thinking chain.<\/p>\n<p>For this reason, the Ziyi-o1 open-source model opts for a smaller parameter scale base model capable of single-card deployments and stronger mathematical capabilities, and is further lightweight enough to be able to run on consumer-grade graphics cards, providing model quality that rivals the quality of cloud-based deployments.<\/p>\n<p>NetEase said that in the scale of \"compression\" at the same time, ZiYu-o1 adopts the thinking chain technology, to create the first output step-by-step lectures thinking chain model, with 14B small parameter scale can reproduce the OpenAI o1 single model reasoning ability.<\/p>","protected":false},"excerpt":{"rendered":"<p>On January 22, the Internet announced today the launch of the country's first output step-by-step reasoning model, \"Sub-o1\", and the official open source. According to the official presentation, sub---o1 is a 14B lightweight single model that supports deployment on a consumer scale graphic card, employing a think-chain technique that provides a detailed problem-solving process with a strong logic and reasoning capability, achieves greater problem-solving accuracy and provides Chinese logic. The ease of the web indicates that there are few open-source models currently available for application and that the size of the parameters makes it impossible to operate on low-visibility consumer-level graphic cards, even if low-bit quantification techniques are used to enable them to be deployed on single cards, but this in turn creates uncertainty in the operation of long-term thinking chains. To this end, the sub-o1 open source model selects the base model of a smaller parameter size that allows<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[5023,5613],"collection":[],"class_list":["post-27595","post","type-post","status-publish","format-standard","hentry","category-news","tag-5023","tag-5613"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/27595","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=27595"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/27595\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=27595"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=27595"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=27595"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=27595"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}