{"id":17395,"date":"2024-08-07T09:21:37","date_gmt":"2024-08-07T01:21:37","guid":{"rendered":"https:\/\/www.1ai.net\/?p=17395"},"modified":"2024-08-07T09:21:47","modified_gmt":"2024-08-07T01:21:47","slug":"%e6%99%ba%e8%b0%b1ai%e5%ae%a3%e5%b8%83%e5%bc%80%e6%ba%90%e3%80%8c%e6%b8%85%e5%bd%b1%e3%80%8d%e5%90%8c%e6%ba%90%e8%a7%86%e9%a2%91%e7%94%9f%e6%88%90%e6%a8%a1%e5%9e%8b-cogvideox","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/17395.html","title":{"rendered":"Zhipu AI announces the open source of &quot;Qingying&quot; homologous video generation model - CogVideoX"},"content":{"rendered":"<p data-vmark=\"b255\">Smart Spectrum AI today announced that it will<strong>vs.\"<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e6%b8%85%e5%bd%b1\" title=\"[Sees articles with [Small] labels]\" target=\"_blank\" >Qingying<\/a>\"homologous<\/strong>of<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e8%a7%86%e9%a2%91%e7%94%9f%e6%88%90%e6%a8%a1%e5%9e%8b\" title=\"_Other Organiser\" target=\"_blank\" >Video Generation Model<\/a> \u2014\u2014<a href=\"https:\/\/www.1ai.net\/en\/tag\/cogvideox\" title=\"[See articles with [CogVideoX] label]\" target=\"_blank\" >CogVideoX<\/a> <a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%bc%80%e6%ba%90\" title=\"[View articles tagged with [open source]]\" target=\"_blank\" >Open Source<\/a>.<\/p>\n<p data-vmark=\"9732\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-17396\" title=\"31e7c0de-27ce-487b-90a2-5ff397b831b6\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2024\/08\/31e7c0de-27ce-487b-90a2-5ff397b831b6.jpg\" alt=\"31e7c0de-27ce-487b-90a2-5ff397b831b6\" width=\"1080\" height=\"459\" \/><\/p>\n<p data-vmark=\"a7a7\">The CogVideoX open source model is described as containing several models of different sizes and dimensions.<strong>Currently open-sourcing CogVideoX-2B.<\/strong>It requires 18GB of video memory for inference at FP-16 precision and 40GB for fine-tuning, which means that<strong>Reasoning with a single 4090 graphics card<\/strong>but (not)<strong>Fine-tuning with a single A6000 graphics card<\/strong>.<\/p>\n<p data-vmark=\"e395\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-17397\" title=\"98827008-680c-4027-b6d4-be7c206edc0d\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2024\/08\/98827008-680c-4027-b6d4-be7c206edc0d.jpg\" alt=\"98827008-680c-4027-b6d4-be7c206edc0d\" width=\"1080\" height=\"435\" \/><\/p>\n<p data-vmark=\"fce3\">CogVideoX-2B has a cue word limit of 226 tokens.<strong>Video length is 6 seconds<\/strong>The frame rate is 8 frames per second and the video resolution is 720*480.<\/p>\n<p data-vmark=\"1999\">Officials said,<strong>Models with higher performance and higher number of parameters are on the way!<\/strong>Please stay tuned and look forward to it.<\/p>\n<p data-vmark=\"4f8e\">Attached related links:<\/p>\n<ul class=\"medium-size list-paddingleft-2\">\n<li>\n<p data-vmark=\"7b7d\">Code Repository:<a href=\"https:\/\/github.com\/THUDM\/CogVideo\" target=\"_blank\" rel=\"noopener\"><span class=\"link-text-start-with-http\">https:\/\/github.com\/THUDM\/CogVideo<\/span><\/a><\/p>\n<\/li>\n<li>\n<p data-vmark=\"dc4f\">Model Download:<a href=\"https:\/\/huggingface.co\/THUDM\/CogVideoX-2b\" target=\"_blank\" rel=\"noopener\"><span class=\"link-text-start-with-http\">https:\/\/huggingface.co\/THUDM\/CogVideoX-2b<\/span><\/a><\/p>\n<\/li>\n<li>\n<p data-vmark=\"cfca\">Technical report:<a href=\"https:\/\/github.com\/THUDM\/CogVideo\/blob\/main\/resources\/CogVideoX.pdf\" target=\"_blank\" rel=\"noopener\"><span class=\"link-text-start-with-http\">https:\/\/github.com\/THUDM\/CogVideo\/blob\/main\/resources\/CogVideoX.pdf<\/span><\/a><\/p>\n<\/li>\n<\/ul>","protected":false},"excerpt":{"rendered":"<p>Wisdom Spectrum AI announced today that it has open-sourced CogVideoX, a video generation model that shares the same origin as \"Clear Shadow\". According to the introduction, the CogVideoX open source model contains several models of different sizes, and will now open source CogVideoX-2B, which requires 18GB of video memory for inference at FP-16 precision, and 40GB of video memory for fine-tuning, which means that inference can be performed on a single 4090 VGA, and fine-tuning can be accomplished on a single A6000 VGA. CogVideoX-2B has a cue word limit of 226 tokens, a video length of 6 seconds, a frame rate of 8 frames per second, and a video resolution of 720*480. Officially, the more powerful cue words with a larger number of parameters will be able to be used with the CogVideoX-2B.<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[3731,219,379,3732,460],"collection":[],"class_list":["post-17395","post","type-post","status-publish","format-standard","hentry","category-news","tag-cogvideox","tag-219","tag-ai","tag-3732","tag-460"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/17395","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=17395"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/17395\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=17395"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=17395"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=17395"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=17395"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}