Google Releases New Large Language Model: Zero Shot Generated 10-Second Video to SOTA!Google's latest large language model VideoPoet can generate videos based on text content, effects

Google Releases New Large Language Model: Zero-Shot Generation of 10-Second Video to SOTA!

Google's latest large language model, VideoPoet, can generate videos based on text content with stable effects, realistic movements, and high definition. It can perform operations such as text-to-video, image-to-video, video editing, stylization, picture supplement, text-to-video task, video output length adjustment, video style adjustment, video editing, picture supplement, video length adjustment, and video soundtrack, and can generate 1-minute long videos. In addition, VideoPoet can be improved by leveraging existing large models and employing MAGVIT V2 and SoundStream to process video images and audio representations.