知识卡片
竞争成绩不能证明另一任务能力
毒性多标签AUC评估与德英翻译属于不同任务,不能用前者0.989结果来讨论翻译错误或语法能力。本章结尾混用两任务叙述,理解时应各自保留数据、指标和输出对象,不外推为通用理解或推理能力。
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
原书出处 · 1 处
- · 查看原书依据
Now, that’s looking better. A 0.989 accuracy puts us in contention with the top challenge solutions of early 2018. And perhaps, you think that 98.9% accuracy may be a little too good to be true. You’d be right. Someone fluent in German would need to dig into several of the translations to find all the translation errors your model is making. And the false negatives—test
参考来源
- 位置:《Natural Language Processing in Action, Second Edition》第09章《可堆叠的Transformer》(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-9.html)
- 结论依据:摘录中关于竞争成绩不能证明另一任务能力的定义、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:Now, that’s looking better. A 0.989 accuracy puts us in contention with the top challenge solutions of early 2018. And perhaps, you think that 98.9% accuracy may be a little too good to be true. You’d be right. Someone fluent in German would need to dig into several of the translations to find all the translation errors your model is making. And the false negatives—test