知识卡片

普通读书笔记卡 · 9264 · 生成模型:GPT-6

实时推理不等于部署时在线训练

理解等级:中级

车辆可实时处理传感器输入而保持模型权重固定。在线更新是另外的训练与发布过程,需明确数据、验证和生效范围;不能把实时运行自动理解为每辆车持续自改并广播全车队。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      The most prevalent application of ML in autonomous vehicles is for self-driving cars. Such vehicles employ an array of sensors, cameras, and lidar to perceive their environment, with ML algorithms processing the data collected from these sensors in real time. Deep learning models are usually trained offline on a collection of real-world and simulated data. During deployment, models can also be actively trained. The changes

    参考来源

    - 位置:《Machine Learning Fundamentals: Concepts, Models, and Applications》第10章《现实应用》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Machine Learning Fundamentals: Concepts, Models, and Applications/_epub-src/OEBPS/xhtml/25_chapter10.xhtml) - 结论依据:摘录中关于实时推理不等于部署时在线训练的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:The most prevalent application of ML in autonomous vehicles is for self-driving cars. Such vehicles employ an array of sensors, cameras, and lidar to perceive their environment, with ML algorithms processing the data collected from these sensors in real time. Deep learning models are usually trained offline on a collection of real-world and simulated data. During deployment, models can also be actively trained. The changes