Nvidia 把世界模型的门槛打下来了——2.6B 参数、单 GPU 可跑、开源，做具身 AI 和机器人仿真的团队可以直接拿来用，省去从头训练的麻烦。
Nvidia 发布了名为 Cosmos 的 2.6B 参数开源世界模型，能够将单张图片、文本提示和轨迹转化为可控的虚拟世界。该模型可在单张 GPU（如 RTX 5090 或 H100）上运行，大幅降低了世界模型的使用门槛。代码和论文已分别发布在 GitHub 和 arXiv 上。该模型适用于具身 AI、机器人研究和仿真等场景，让更多研究者和开发者能够探索世界模型的应用。
Woow Nvidia has just released a 2.6B open-source world model 🔥 You can turn a single image, text p...
Woow Nvidia has just released a 2.6B open-source world model 🔥 You can turn a single image, text prompt and trajectory into controllable worlds... And on a single GPU! - Code available on GitHub - Paper as well on arxiv You can use it for many things like embodied AI and robotics research, simulations, etc. Because it can run on a single GPU (like an RTX 5090 or H100) it makes world models accessible to basically everyone! Your browser does not support the video tag. 🔗 View on Twitter 💬 17 🔄 84 ❤️ 593 👀 37533 📊 181 ⚡