知识卡片
预切分限制后续合并范围
预切分先把文本分成若干片段,再在片段内做子词切分。按空格、标点或数字分割的规则随模型而变,可能保留连续空格或词内标点;不能把某个Whitespace正则当所有tokenizer共同规则。
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
- · 词表与分词
原书出处 · 1 处
- · 查看原书依据
Pre-Tokenization Before we run the tokenizer on the text, we can optionally perform a pre-tokenization step. As mentioned earlier, most tokenizers today employ subword tokenization. A common step is to first perform word tokenization and then feed the output of it to the subword tokenization algorithm. This step is called pre-tokenization. Pre-tokenization is a relatively easy step in English compared to many other languages, since
参考来源
- 位置:《Designing Large Language Model Applications》第03章《词表与分词》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Designing Large Language Model Applications/_epub-src/OEBPS/ch03.html)
- 结论依据:摘录中关于预切分限制后续合并范围的定义、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:Pre-Tokenization Before we run the tokenizer on the text, we can optionally perform a pre-tokenization step. As mentioned earlier, most tokenizers today employ subword tokenization. A common step is to first perform word tokenization and then feed the output of it to the subword tokenization algorithm. This step is called pre-tokenization. Pre-tokenization is a relatively easy step in English compared to many other languages, since