知识卡片

普通读书笔记卡 · 6701.a · 生成模型:GPT-6

全词表预测成本随词表扩大

理解等级:中级

示例输出层将嵌入扩展到词表每项,再用交叉熵训练,具有全词表计算成本。它展示基本机制,不能与负采样训练在成本或目标上完全等同;词表、维数和批量会共同影响内存及时间。

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

      self.expand = nn.Linear(embedding_size, vocab_size, bias=False) def forward(self, input): hidden = self.embed(input) #3 logits = self.expand(hidden) #4 return logits #1 100 is small but usable for many problems, while 300 is more typical. #2 Initializes the layers of your network only once, when you instantiate the Word2Vec object #3 The hidden layer embeds (encodes) the statistics of word usage in a lower-dimensional vector. #4 The output

    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:摘录中关于全词表预测成本随词表扩大的定义、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:self.expand = nn.Linear(embedding_size, vocab_size, bias=False) def forward(self, input): hidden = self.embed(input) #3 logits = self.expand(hidden) #4 return logits #1 100 is small but usable for many problems, while 300 is more typical. #2 Initializes the layers of your network only once, when you instantiate the Word2Vec object #3 The hidden layer embeds (encodes) the statistics of word usage in a lower-dimensional vector. #4 The output