知识卡片
模型变大仍需按目标样本验证
较大模型在书中例子改正了部分人名和组织标签,不表示全部数据都提高。模型、分词、配置与输入变更也会改变token位置;应使用固定目标集合测边界和类型,并保存错误去向。
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
A larger spaCy model should be able to improve your accuracy a little bit, especially for words that aren’t very common in the datasets used to train spaCy: >>> nlp = spacy.load('en_core_web_lg') >>> doc = nlp(text) >>> doc2df(doc) TOK POS TAG ENT_TYPE DEP 0 Timnit PROPN NNP PERSON compound 1 Gebru PROPN NNP PERSON nsubj 2 ( PUNCT -LRBpunct 3 Amharic PROPN NNP NORP appos
参考来源
- 位置:《Natural Language Processing in Action, Second Edition》第11章《信息抽取与知识图谱》(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-11.html)
- 结论依据:摘录中关于模型变大仍需按目标样本验证的定义、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:A larger spaCy model should be able to improve your accuracy a little bit, especially for words that aren’t very common in the datasets used to train spaCy: >>> nlp = spacy.load('en_core_web_lg') >>> doc = nlp(text) >>> doc2df(doc) TOK POS TAG ENT_TYPE DEP 0 Timnit PROPN NNP PERSON compound 1 Gebru PROPN NNP PERSON nsubj 2 ( PUNCT -LRBpunct 3 Amharic PROPN NNP NORP appos