← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6706 · 生成模型:GPT-6

稠密存储仍有实际类型和副本成本

理解等级:中级

300万词×300维若用float32,仅向量数值区就需36亿字节,另有词键、范数及对象成本。内存映射可减少一次性驻留,但HDF5或DataFrame不天然都按同样方式映射;需检查加载器和后续操作是否物化。

这张卡片看不懂?

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

      Warning The Google News Word2Vec model is huge, boasting 3 million words with 300 vector dimensions each. The complete word vector model requires 3 GB of RAM. The KeyedVectors object in gensim now holds a table of 3 million Word2Vec vectors. We loaded these vectors from a file created by Google to store a Word2Vec model they trained on a large corpus based on Google

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第06章《用词嵌入推理》(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-6.html) - 结论依据:摘录中关于稠密存储仍有实际类型和副本成本的定义、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:Warning The Google News Word2Vec model is huge, boasting 3 million words with 300 vector dimensions each. The complete word vector model requires 3 GB of RAM. The KeyedVectors object in gensim now holds a table of 3 million Word2Vec vectors. We loaded these vectors from a file created by Google to store a Word2Vec model they trained on a large corpus based on Google