字节跳动否认推出Boximator视频生成模型。近日，有消息称字节跳动推出了名为Boximator的视频生成模型，但该公司回应称，Boximator实际上是一个关于视频生成领域控制对象运动的技术研究项目，尚未达到产品化阶段。相比于Gen-2、Pink1.0等模型，Boximator被认为能够通过文本精细控制视频中人物或物体的动作。然而，字节跳动表示，Boximator在画面质量、保真率、视频时长等方面仍与国外领先模型存在较大差距。这一声明澄清了外界对于字节跳动在这一领域技术进展的误解。
Title: ByteDance Denies Releasing Boximator Video Generation Model
Keywords: ByteDance, Boximator, Video Generation
News content:
ByteDance has denied the release of the Boximator video generation model. Recent reports suggested that ByteDance had launched a video generation model called Boximator, but the company clarified that Boximator is actually a research project in the field of video generation technology, focusing on controlling the movement of objects in videos through text inputs. However, ByteDance stated that Boximator is not yet at a product level and that there are significant gaps in terms of image quality, fidelity, and video length compared to leading international models. This statement clears up misunderstandings about ByteDance’s progress in this area of technology.
【来源】https://finance.sina.com.cn/tech/shenji/2024-02-20/doc-inairynk7158099.shtml
Views: 4