{"id":2679,"date":"2024-01-08T10:27:16","date_gmt":"2024-01-08T02:27:16","guid":{"rendered":"https:\/\/www.1ai.net\/?p=2679"},"modified":"2024-01-08T10:27:16","modified_gmt":"2024-01-08T02:27:16","slug":"meta%e6%8e%a8ai%e8%a7%86%e9%a2%91%e6%a8%a1%e5%9e%8bfairy%ef%bc%8c%e8%bd%bb%e6%9d%be%e6%9b%bf%e6%8d%a2%e8%a7%86%e9%a2%91%e4%ba%ba%e7%89%a9%e3%80%81%e6%94%b9%e5%8f%98%e9%a3%8e%e6%a0%bc","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/2679.html","title":{"rendered":"Meta launches AI video model Fairy, which can easily replace video characters and change styles"},"content":{"rendered":"

Meta<\/a> The GenAI team launched a Fairy<\/a> The team demonstrated Fairy\u2019s performance in several applications, including character\/object replacement, stylization, and long-form video generation.<\/p>\n

For example, a simple text prompt, such as "in the style of Van Gogh," is enough to edit the source video. For example, the text command "turn into a snowman" turns the astronaut in the video into a snowman.<\/p>\n

\"\"<\/p>\n

\"\"<\/p>\n

Visual coherence in Fairy is a particularly challenging problem, as there are countless ways to modify a given image based on the same cue. Fairy uses cross-frame attention, a mechanism that implicitly propagates diffuse features, ensuring superior temporal coherence and high-fidelity synthesis.<\/p>\n

\"\"<\/p>\n

The model can generate a video of 512 x 384 pixels, 120 frames (4 seconds under 30fps) in only 14 seconds, at least 44 times faster than the previous model. Like Meta's Emu video model, Faery is based on a proliferation model for image processing, enhanced by video editing\u3002<\/p>\n

Fairy processes all frames of the source video without temporal downsampling or frame interpolation, and maintains the aspect ratio of the horizontal output video at 512. When tested with six A100 GPUs, Fairy was able to render a 27-second video in 71.89 seconds with high visual consistency.<\/p>\n

Fairy\u2019s performance was tested in an extensive user study with 1,000 generated samples. Both human judgment and quantitative metrics confirmed that Fairy outperformed Rerender, TokenFlow, and Gen-1.<\/p>\n

However, the model currently has problems handling dynamic environmental effects like rain, fire, or lightning, which either don't fit well into the overall scene or produce visual errors.<\/p>\n

Despite these issues, the research team believes their work represents a significant advance in the field of AI video editing, with a transformative approach to temporally consistent and high-quality video synthesis.<\/p>","protected":false},"excerpt":{"rendered":"

Meta's GenAI team has unveiled a video-to-video synthesis model called Fairy that is faster and more temporally consistent than existing models. The team demonstrated Fairy's performance in several applications, including character\/object replacement, stylization and long form video generation. For example, a simple text prompt such as \"Imitate Van Gogh style\" was sufficient to edit the source video. For example, the text command \"turn into snowman\" turns the astronaut in the video into a snowman. Visual coherence in Fairy is a particularly challenging problem because there are countless ways to modify a given image based on the same cue; Fairy uses cross-frame focusing, a mechanism that implicitly propagates diffusion features, to ensure superior temporal coherence.<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[868,867,297],"collection":[],"class_list":["post-2679","post","type-post","status-publish","format-standard","hentry","category-news","tag-ai","tag-fairy","tag-meta"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/2679","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=2679"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/2679\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=2679"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=2679"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=2679"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=2679"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}