Today, Beijing-based tech company Shengshu Technology announced the global launch of the official website for their AI video generation model, Vidu. Vidu is an innovative model that harnesses artificial intelligence to generate videos from text or images.
In April of this year, Shengshu Technology in collaboration with Tsinghua University unveiled China's first video large model, "Vidu," marking a significant step forward in video generation technology in China.
"Vidu" incorporates the team's original U-ViT architecture, which integrates Diffusion and Transformer technologies. This innovative video large model can quickly generate 16-second, 1080P high-definition videos, showcasing high imagination and creativity while simulating a realistic physical world. Its multi-camera generation capability and temporal consistency are notable features of "Vidu."
Since its release, "Vidu" has achieved significant breakthroughs globally, reaching top-tier international performance levels and continuously undergoing iterations and optimizations. This achievement is attributed to the team's deep expertise in Bayesian machine learning and multi-modal large models, as well as several original contributions.
Leveraging a profound understanding of the U-ViT architecture and extensive engineering and data experience, the team swiftly overcame key technical challenges in long video representation and processing, successfully developing the "Vidu" video large model. "Vidu" has made notable advancements in enhancing video coherence and dynamism, propelling the development of video processing technology.
Experience it here: https://top.aibase.com/tool/viduguanwang