← 返回起始笔记位置

知识卡片

普通读书笔记卡 · 8980.c · 生成模型:GPT-6

无条件生成仍受训练分布约束

理解等级:中级

无条件模型描述p(x),不接收每次生成所需的外部描述;条件模型描述p(x|y)。无条件不意味着没有数据、先验或模型假设,训练内容和结构仍规定会产生什么。两种模型的区别是条件输入的角色,而非输出是否随机。

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

      We call this an unconditional generative model because it is model of the unconditional distribution p(X). Generally we will refer to unconditional generative models simply as generative models and use the term conditional generative model for a model of any conditional distribution p(X | Y). Conditional generative models will be the focus of chapter 34; in the present chapter we will restrict our attention, from

    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:摘录中关于无条件生成仍受训练分布约束的定义、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:We call this an unconditional generative model because it is model of the unconditional distribution p(X). Generally we will refer to unconditional generative models simply as generative models and use the term conditional generative model for a model of any conditional distribution p(X | Y). Conditional generative models will be the focus of chapter 34; in the present chapter we will restrict our attention, from