AWS 与 NVIDIA 合作：在 SageMaker HyperPod 上构建物理 AI 模型工厂
原标题：Build a Physical AI model factory with NVIDIA Cosmos 3 on SageMaker HyperPod
AI 摘要
AWS 博客介绍了如何利用 NVIDIA Cosmos 3 在 SageMaker HyperPod 上构建物理 AI 模型工厂。Cosmos 3 采用混合 Transformer 架构，统一处理视频、图像、动作和声音，支持前向动力学、逆动力学和动作策略三种模式。该设计使数据生成、后训练和评估可在同一 GPU 集群上运行，提高资源利用率。文章提供了端到端示例，并强调 GPU 有效吞吐是成本控制的关键。
正文节选
Build a Physical AI model factory with NVIDIA Cosmos 3 on SageMaker HyperPod A Physical AI system, such as a robot or autonomous vehicle (AV) that translates real-world data into physical actions, can’t be built in a single training job. Instead, it takes a continuous pipeline: a loop of generating synthetic data, post-training perception and policy models, so the system understands its surroundings and can act, and evaluating both in closed-loop simulation. Running that pipeline continuously is