知识卡片

普通读书笔记卡 · 6511.f · 生成模型:GPT-6

负号把奖励最大化转成最小化

理解等级:中级

最小化 -mean(A_i log pθ(y_i|x)) 的局部梯度,会提升正优势轨迹的对数概率并压低负优势轨迹的对数概率。该解释针对当前样本和目标;共享参数、采样方差及其他优化项意味着实际全局能力变化仍需评估。

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 list of the rollout logprobs from the previous section into a PyTorch tensor. Then, we compute the policy gradient loss by multiplying each rollout's sequence-level logprob by its corresponding advantage, taking the mean across rollouts, and applying a negative sign. Regarding the negative sign, because PyTorch optimizers are defined to minimize a loss, objectives that are naturally written as maximization problems must be sign-flipped.

    参考来源

    - 位置:《Build a Reasoning Model (From Scratch)》第06章《强化学习训练推理模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Build a Reasoning Model (From Scratch)/_epub-src/OEBPS/Text/chapter-6.html) - 结论依据:摘录中关于负号把奖励最大化转成最小化的定义、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 list of the rollout logprobs from the previous section into a PyTorch tensor. Then, we compute the policy gradient loss by multiplying each rollout's sequence-level logprob by its corresponding advantage, taking the mean across rollouts, and applying a negative sign. Regarding the negative sign, because PyTorch optimizers are defined to minimize a loss, objectives that are naturally written as maximization problems must be sign-flipped.