← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9385.a · 生成模型:GPT-6

超参包括模型之外的预处理选择

理解等级:中级

分词、停用词、词表阈值和归一化会与模型容量共同影响结果,应纳入实验配置。训练学得权重与外部选择的超参不同;调整预处理时需按完整流程重新验证,避免只报告最终模型名称。

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

      Hyperparameters Hyperparameters are all the values that determine the performance of your pipeline, including the model type and how it is configured. This can include things like how many neurons and layers are in a neural network or the value of alpha in an sklearn.linear_model.Ridge regressor. Hyperparameters also include the values that govern any preprocessing steps, like the tokenizer type, a list of ignored words,

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》附录D《附录D:机器学习工具与技术》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/appendix-d.html) - 结论依据:摘录中关于超参包括模型之外的预处理选择的定义、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:Hyperparameters Hyperparameters are all the values that determine the performance of your pipeline, including the model type and how it is configured. This can include things like how many neurons and layers are in a neural network or the value of alpha in an sklearn.linear_model.Ridge regressor. Hyperparameters also include the values that govern any preprocessing steps, like the tokenizer type, a list of ignored words,