Shenzhen Yuanxiang Information Technology Co., Ltd. recently released MotionGen, the country’s first physics-based 3D action generation model. This model uses technologies such as large models, physical simulation, and reinforcement learning to generate realistic and smooth 3D actions through simple text instructions. . This marks a major breakthrough in China's 3D AIGC field, and also brings higher creative efficiency and freedom to industries such as animation, games, movies, and virtual reality. MotionGen's low-entry feature allows even creators with zero experience to easily get started and create high-quality animated content. Its core technology, the MotionGenGPT algorithm, can generate complex 3D movements without manual intervention and can be applied to the skeleton driving of any character.
The MotionGen model uses precise text understanding capabilities to seamlessly convert text instructions into corresponding 3D actions, supporting various action ideas from basic walking to complex body movements. Currently, the MotionGen model is in the testing phase, and users can apply for a free trial. Yuanxiang XVERSE is committed to promoting the development of the AI+3D field, with the ultimate goal of allowing everyone to freely define their own virtual world. The emergence of MotionGen will undoubtedly profoundly change the way of 3D content creation and promote the process of virtual world construction. Experience address: www.MotionGen.cn
It is believed that the launch of the MotionGen model will accelerate the popularity of 3D content creation and bring more exciting content to the virtual world. In the future, with the continuous iteration and improvement of technology, MotionGen will have more powerful functions and wider application scenarios, creating a richer experience for users.