知识卡片

普通读书笔记卡 · 8980 · 生成模型:GPT-6

随机输入表达生成多样性

理解等级:中级

同一个类别或描述可以对应许多图像,用随机z和确定函数g(z,y)即可表达这种一对多生成。确定的是给定全部输入后的运算,不是给定y后的唯一图像。随机分布和映射共同决定输出分布,分类器的反方向类比并不构成数学逆函数。

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

      One way to make a function stochastic is to make it a deterministic function of a stochastic input, and this is how most of the generative models in this chapter will work. We define g as a function, called a generator, that takes a randomized vector z as input along with the label/description y, and produces an image x as output, that is, g :

    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:摘录中关于随机输入表达生成多样性的定义、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:One way to make a function stochastic is to make it a deterministic function of a stochastic input, and this is how most of the generative models in this chapter will work. We define g as a function, called a generator, that takes a randomized vector z as input along with the label/description y, and produces an image x as output, that is, g :