China's First Sora-Level Text-to-Video Large Model Vidu Unveiled
China unveils its first Sora-level text-to-video large model, Vidu. Vidu can generate high-definition videos in just one click. Vidu offers Chinese localisation features, setting it apart from its competitors.
In a groundbreaking development, China has unveiled its first Sora-level text-to-video large model, Vidu. The unveiling took place at the prestigious Zhongguancun Forum on Saturday, marking a significant milestone in the global artificial intelligence competition.
Vidu, developed by Chinese AI firm Shengshu Technology in collaboration with Tsinghua University, promises to revolutionise the creation of high-definition videos. According to China Daily, this cutting-edge model can generate a 16-second video with a resolution of 1080p in just one click.
What sets Vidu apart is its extended duration, exceptional consistency, and dynamic capabilities, making it the first of its kind in China. The developers claim that Vidu is "very close to" the level of Sora, a generative AI model developed by the renowned American company OpenAI earlier this year. Sora has gained widespread acclaim for its ability to construct realistic and imaginative scenes based on text instructions.
While Sora has captivated the tech world with its prowess, Vidu offers a unique advantage by understanding and generating Chinese elements, such as the iconic panda and dragon. This localisation feature makes Vidu particularly appealing for Chinese users and opens up new possibilities for creative content generation.
Interestingly, the core architecture of Vidu was initiated in September 2022, predating Sora's adoption of its architecture. This demonstrates China's commitment to pushing the boundaries of AI technology and staying at the forefront of innovation.
China unveils its first Sora-level text-to-video large model, Vidu
Vidu can generate high-definition videos in just one click
Vidu offers Chinese localisation features, setting it apart from its competitors
Source: CHINA DAILY