share_log

快手可灵大模型再进化 图生视频及视频续写功能发布

Kuaishou's Ke Ling AI model has evolved again with the release of the video creation and continuation functions.

Gelonghui Finance ·  Jun 21 15:08
On June 21st, Kuaishou's video generation model "Keling" attracted widespread attention both domestically and abroad due to its amazing effects. On June 21st, Ke Ling was once again evolved to officially launch the graphic-to-video function, which supports converting any static image into a 5-second vivid video with different motion effects generated by the input text of creators. Ke Ling also simultaneously released the industry-leading video continuation function, supporting one-click continuation and continuous multi-time continuation of the generated video. One continuation can extend the video for about 5 seconds and the longest can generate videos of about 3 minutes, further demonstrating the model's strong imagination and controllability. It is reported that Kuaishou's "Ke Ling" video generation model was officially released on June 6th, and not only is its effect on par with Sora, but it also opened for testing in the Kuaishou Video app, with about 140,000 people queuing up to apply. Today, Kuaishou also released the latest functions of Ke Ling graphic-to-video and video continuation at the computer vision conference CVPR2024, which received strong feedback.

The translation is provided by third-party software.


The above content is for informational or educational purposes only and does not constitute any investment advice related to Futu. Although we strive to ensure the truthfulness, accuracy, and originality of all such content, we cannot guarantee it.
    Write a comment