知识卡片
编码器和分类头可独立发布
大嵌入器与小分类头有不同体积和更新频率,拆开可复用公共表示并更换领域标签模型。分类头仍依嵌入维度、模型版本及标签顺序,不能任意混用;更小制品也不自动降低整条服务延迟。
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
The entire pipeline has been pickled with the joblib library and stored in a public object storage bucket. Because this pipeline contains the BERT model to create the embeddings for this classifier, it takes more than 80 MB of storage for the joblib (PKL) file. This large model file will be automatically downloaded the first time you run the classifier in a new environment so
参考来源
- 位置:《Natural Language Processing in Action, Second Edition》附录E《附录E:容器化微服务部署》(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-e.html)
- 结论依据:摘录中关于编码器和分类头可独立发布的定义、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:The entire pipeline has been pickled with the joblib library and stored in a public object storage bucket. Because this pipeline contains the BERT model to create the embeddings for this classifier, it takes more than 80 MB of storage for the joblib (PKL) file. This large model file will be automatically downloaded the first time you run the classifier in a new environment so