← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 8988.f · 生成模型:GPT-6

生成能力需要多个维度判断

理解等级:中级

可评分、可抽样、可用低维变量控制和容易反演,是不同接口能力;速度、逼真、多样性与泛化又是不同质量指标。书中的表格按典型使用方式概括模型,应作为理解取舍的起点,而非断言某类模型永久无法实现表中未勾选能力。

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

      Some of the generative models we will describe in this chapter and the next are categorized along these dimensions in table 37.1. Note that this is a rough categorization, meant to reflect the most straightforward usage of these models. With additional effort some of the ✗'s can be converted to ✓'s; for example, one can do additional inference to sample from a density model, or

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第32章《生成模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)/_epub-src/OPS/xhtml/c032.xhtml) - 结论依据:摘录中关于生成能力需要多个维度判断的定义、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:Some of the generative models we will describe in this chapter and the next are categorized along these dimensions in table 37.1. Note that this is a rough categorization, meant to reflect the most straightforward usage of these models. With additional effort some of the ✗'s can be converted to ✓'s; for example, one can do additional inference to sample from a density model, or