知识卡片
词元编号可避免显式存全部零
用一百万词表为1.575亿个位置显式存一字节独热元素,需要157.5万亿字节;按位紧凑存储约19.69万亿字节。更常用稀疏结构或词元编号保存身份,查嵌入等操作不必先建立全语料密集独热矩阵。
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
- · 思维的词元
原书出处 · 1 处
- · 查看原书依据
Let’s run through the math to give you an appreciation for just how big and unwieldy these “piano rolls” are. In most cases, the vocabulary of tokens you’ll use in an NLP pipeline will be much more than 10,000 or 20,000 tokens—sometimes, it can be hundreds of thousands or even millions of tokens. Let’s assume you have a million tokens in your NLP pipeline vocabulary,
参考来源
- 位置:《Natural Language Processing in Action, Second Edition》第02章《思维的词元》(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-2.html)
- 结论依据:摘录中关于词元编号可避免显式存全部零的定义、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:Let’s run through the math to give you an appreciation for just how big and unwieldy these “piano rolls” are. In most cases, the vocabulary of tokens you’ll use in an NLP pipeline will be much more than 10,000 or 20,000 tokens—sometimes, it can be hundreds of thousands or even millions of tokens. Let’s assume you have a million tokens in your NLP pipeline vocabulary,