ByteDance responded to rumors about its launch of a Chinese version of the sora** model, saying that there is no perfect product to land. Boximator, which is rumored to be a new ** model developed by ByteDance, is described as a subversive model that can be used to generate the movements of people or objects in ** through precise control of text. However, ByteDance denied this, saying that Boximator was only a research project on technical methods to control the movement of objects in the generation domain.
ByteDance said that Boximator is currently unable to land as a perfect product, and compared with the leading foreign generation models, there is still a big gap in terms of picture quality, fidelity, and duration. This suggests that while the boximator may have some innovative technical features, it is not currently up to the standard for practical application.
As a text-to-text model capable of generating complex scenes with multiple characters, specific movement types, and precise subject and background details, the emergence of SORA has undoubtedly brought new possibilities to the field of generation. However, ByteDance's response suggests that while they may be exploring similar technologies, they have not yet been able to reach the level of effectiveness and application of Sora-like models.
According to the data of business investigation, Douyin Vision*** currently has 7889 patent information, of which 868 are patents. ByteDance said this time that they are actively exploring in the field of first-class generation technology, but in order to meet the standards of practical application, further improvements and promotions are needed in many aspects.