Event ·Multi-source

NVIDIA开源PixelUMM:一个模型同时支持图文和视频的理解与生成

NVIDIA开源了无编码器多模态模型PixelUMM,一套架构同时处理图像和视频的理解与双向生成。它去掉了传统独立的视觉编码器,把画面切分成像素块直接送入主模型,让输入理解和输出生成共用同一套底层表示。这种设计简化了多模态流程,模型和权重已完整开源。

Current read

NVIDIA把PixelUMM开源了,一个模型搞定图文视频的理解和生成,做法是不用视觉编码器直接吃像素块,搞多模态研究的可以看看源码。

11 sources71° AI heatLast updated 2026/10/5 12:03:38

Evidence

11 sources
Related source 1kimmonismus
View source
Related source 2techcrunch
View source
Related source 3The Business Times: Tech
View source
Related source 4The Straits Times: Business
View source
Related source 5marktechpost
View source
Related source 7Justine Moore
View source
Related source 8Tech in Asia
View source
Related source 9Nathan Lambert
View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.