According to a statement from Shenshu, Vidu can generate a four-second clip in 30 seconds, making it one of the fastest on the market, as other similar tools typically take much longer to generate a video of the same length.
Shengshu exemplifies how China's prestigious Tsinghua University has emerged as a main force behind the country's AI ambitions. Behind Vidu is the company's self-developed architecture, called U-ViT, which was first detailed in a September 2022 research paper written by a team led by Zhu Jun, Shengshu AI's chief scientist and a professor of computer science at Tsinghua University.
The paper's other author, Bao Fan, from Tsinghua University, is now chief technology officer at Shengshu. Shengshu's CEO, Tang Jiayu, is a graduate of Tsinghua University's School of Computer Science and Technology.
In an April interview, Tang told local media that it would be easier for Chinese companies to catch up with Sora than with OpenAI's advanced large-scale language model GPT-4, the underlying technology behind ChatGPT, without providing details.
In addition to converting text and images into video, Vidu has added features to lay the foundation for commercializing the technology, with potential uses in the animation and content industries in mind, Zhang Xudong, product director at Shengshu AI, told The Post in an interview.
The new character-to-video feature allows users to upload an image of a real person or animated character and bring that character to life using simple text prompts.
“In the future [users] You can upload multiple characters, [describe] “Just like in filmmaking, AI will recreate different scenes and act in them,” Zhang said. “Our goal is to integrate AI tools into traditional fields.”
Sheng Shu has raised tens of millions of dollars and its backers include Qiming Venture Partners, search giant Baidu, Alibaba Group Holding Ltd.'s fintech affiliate Ant Group and the Beijing AI Industry Investment Fund. Alibaba owns Post.
