2024年，英伟达（NVIDIA）正式发布其面向智能汽车领域的全新大模型平台——Alpamayo。该平台以秘鲁著名的阿尔帕马约峰命名，象征其在自动驾驶与车载AI领域的高峰追求。Alpamayo基于NVIDIA强大的AI基础设施构建，整合了生成式AI、多模态感知、自然语言交互和端到端驾驶决策能力，旨在为汽车制造商提供一个可定制、可扩展的智能座舱与自动驾驶解决方案。Alpamayo平台支持在车端部署大型语言模型（LLM）和视觉-语言模型（VLM），使车辆能够理解并响应复杂的用户指令，如导航规划、娱乐控制或车辆状态查询。同时，该平台通过NVIDIA DRIVE Thor芯片实现高效推理，兼顾性能与功耗，适用于从L2+到L4级别的自动驾驶系统。此外，Alpamayo还支持持续学习与OTA（空中下载）更新，让车辆随时间不断进化。英伟达表示，Alpamayo将加速汽车行业向软件定义汽车（Software-Defined Vehicle, SDV）转型，并已与多家全球主流车企展开合作。这一平台的推出，标志着AI大模型正从云端走向边缘设备，深入下一代智能出行的核心。
In 2024, NVIDIA officially launched Alpamayo, a new large-model platform tailored for the intelligent automotive sector. Named after Peru’s iconic Alpamayo peak, the platform symbolizes NVIDIA’s ambition to reach new heights in autonomous driving and in-vehicle AI. Built on NVIDIA’s robust AI infrastructure, Alpamayo integrates generative AI, multimodal perception, natural language interaction, and end-to-end driving decision-making capabilities, offering automakers a customizable and scalable solution for both intelligent cockpits and autonomous driving systems.The Alpamayo platform enables on-device deployment of large language models (LLMs) and vision-language models (VLMs), allowing vehicles to understand and respond to complex user commands—such as route planning, entertainment control, or vehicle status inquiries. Powered by the NVIDIA DRIVE Thor chip, the platform delivers high-efficiency inference with balanced performance and power consumption, supporting autonomy levels from L2+ to L4. Moreover, Alpamayo supports continuous learning and over-the-air (OTA) updates, enabling vehicles to evolve over time.NVIDIA states that Alpamayo will accelerate the automotive industry’s shift toward Software-Defined Vehicles (SDVs) and is already collaborating with several leading global automakers. This launch marks a pivotal step in bringing large AI models from the cloud to edge devices, embedding them at the heart of next-generation mobility.
原创文章，作者：admin，如若转载，请注明出处：https://ktez.cn/9050.html