[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-9652a175-6f95-461d-b9d4-946b44d0fccf":3},{"id":4,"title":5,"summary":6,"original_url":7,"source_id":8,"tags":9,"published_at":23,"created_at":24,"modified_at":25,"is_published":26,"publish_type":27,"image_url":13,"view_count":28},"9652a175-6f95-461d-b9d4-946b44d0fccf","NVIDIA × Hugging Face：Isaac GR00T 1.7 进 LeRobot，把 Cosmos 3 留给开源机器人","英伟达与 Hugging Face 在 CoRL 2026 慕尼黑现场宣布:Isaac GR00T 1.7、Isaac Teleop 与 Isaac Sim\u002FLab 全面接入开源机器人框架 LeRobot,Cosmos 3 世界基础模型也被列入近期路线图。这是首个被两端同时承认的「工业级 VLA + 世界模型 + 仿真 + 数据飞轮」端到端开源流水线,直接打通了 300 万英伟达机器人开发者与 1600 万 Hugging Face AI 用户的协作池。GR00T 1.7 作为首款「开源且可商用」的机器人基础模型,采用视觉-语言-动作(VLA)架构,可在 LeRobot 工作流中后训练并部署到 Jetson Thor 边缘硬件;Cosmos 3 则补齐世界模型一环,提供稀缺的真实交互数据合成能力,缓解具身智能最大的数据瓶颈。配合 NVIDIA 已开源下载超 1500 万次的物理 AI 数据集(35 万条轨迹、5700 万次抓取),整个生态不再围绕闭源方案搭栈。对于正被「数据稀缺 + 工具碎片」卡住的具身研究,这意味着开发者第一次可以拿到一条公开、统一、可在 Jetson Thor 上跑通的 VLA 训练-仿真-部署闭环。","https:\u002F\u002Fblogs.nvidia.com\u002Fblog\u002Fhugging-face-lerobot-models-frameworks-open-robotics\u002F","474eef8c-e0c3-46cf-adee-c089558220f9",[10,14,17,20],{"id":11,"name":12,"slug":12,"description":13,"color":13},"499f4b56-819d-49a3-9609-33e775143b86","multimodal",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"8dac812d-3839-4abe-a855-5f56ec9515fd","nvidia",{"id":18,"name":19,"slug":19,"description":13,"color":13},"b9bd9039-fcdb-41a8-b85b-fc1587def2b9","open-source",{"id":21,"name":22,"slug":22,"description":13,"color":13},"ebe5dcd1-46b1-4298-b8c2-8e0e2f456e56","video-generation","2026-07-07T06:00:26Z","2026-07-09T10:09:59.600098Z","2026-07-09T10:09:59.600110Z",true,"agent",2]