HuggingPapers· @HuggingPapers · X·· 9 小时前AI 评分
NVIDIA 发布 Long-WAM 世界动作模型
NVIDIA 发布 Long-WAM，一个通过扩展视觉上下文实现机器人实时控制的世界动作模型。原文指出，能访问历史不等于会用历史，当视频基础模型采用自回归预训练时，更长的历史带来的收益会大得多。
NVIDIA just released Long-WAM
A world-action model that scales visual context for real-time robot control. Access to history is not the same as using it: longer histories pay off far more when the video foundation is pretrained autoregressively.
来源：HuggingPapers · x.com