知识卡片
损失对象状态不等于学习历史
示例分别创建训练与测试BCELoss,是因为各自携带不同长度和顺序的样本权重。损失函数并不因此累积跨轮学习记忆;若没有这类不同配置,是否复用对象取决于实现,而非所有训练测试都必须新建。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 神经网络基础
原书出处 · 1 处
- · 查看原书依据
#1 Loss functions are stateful, so you will need separate instances for running on the test and training datasets. The last thing we need to choose is how to adjust our weights based on the loss—the optimizer algorithm. Remember that our goal is to minimize the loss function. One way to think of this minimization problem is to imagine ourselves on the slope of the
参考来源
- 位置:《Natural Language Processing in Action, Second Edition》第05章《神经网络基础》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/chapter-5.html)
- 结论依据:摘录中关于损失对象状态不等于学习历史的定义、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:#1 Loss functions are stateful, so you will need separate instances for running on the test and training datasets. The last thing we need to choose is how to adjust our weights based on the loss—the optimizer algorithm. Remember that our goal is to minimize the loss function. One way to think of this minimization problem is to imagine ourselves on the slope of the