← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 8981.c · 生成模型:GPT-6

生成器与密度函数各有接口

理解等级:中级

生成器接随机输入产出样本,密度函数对给定数据求概率密度,能量函数提供未归一化分数。模型可能提供其中一种或多种,不能因能生成就假设能高效算似然,也不能因能评分就假设已拥有便宜可靠的采样器。

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 generative models not only produce generators but also yield a probability density function pθ, fit to the training data. This density function may play a role in training the generator (e.g., we want the generator to produce samples from pθ), or the density function may be the goal itself (e.g., we want to be able to estimate the probability of datapoints in order to

    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:摘录中关于生成器与密度函数各有接口的定义、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 generative models not only produce generators but also yield a probability density function pθ, fit to the training data. This density function may play a role in training the generator (e.g., we want the generator to produce samples from pθ), or the density function may be the goal itself (e.g., we want to be able to estimate the probability of datapoints in order to