China's First Video Large Model Vidu to Undergo Major Updates
TapTechNews, June 7 - According to the Thursday report of 'Titanium Media AGI', the video large model Vidu, jointly released by Shengshu Technology and Tsinghua University, the first of its kind in China with long duration, high consistency, and high dynamics, will complete three major updates in the near future to achieve a'significant technical iteration'.
The main update contents are as follows:
Supports one-click generation of 32-second videos.
Supports audio-video synthesis, 'videos with sound' (Text-2-Audio).
Supports 4D generation, which can generate 4D content with temporal and spatial consistency from a single video.
TapTechNews attaches partial video previews:
According to the introduction, Vidu is the first video large model in the world to achieve a major breakthrough after the release of Sora, and it is still accelerating iterative improvement.
According to TapTechNews' previous reports, the Vidu model fuses Diffusion and Transformer, and pioneered the creation of U-ViT. When it was released in April, it supported one-click generation of high-definition video content with a length of up to 16 seconds and a resolution of up to 1080P. The official said that Vidu can not only simulate the real physical world but also has rich imagination and features such as multi-lens generation and high temporal and spatial consistency.
Related reading:
'China's First Sora-level Model Vidu Released: Can Generate Videos with a Maximum of 16 Seconds and a Maximum of 1080P'