知识卡片

普通读书笔记卡 · 6634 · 生成模型:GPT-6

编码器与解码器按可见性分工

理解等级:中级

编码器通常为输入位置建立双向表示;编码器解码器分开处理条件输入与自回归输出;仅解码器用相应遮罩在同一序列组织上下文和续写。骨架、遮罩与训练目标要共同看,不能只凭模型名判断最适任务。

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

      Encoder-only Encoder-decoder Decoder-only Let’s look at each of these in detail. Figure 4-5 depicts encoder-only, encoder-decoder, and decoder-only architectures. Figure 4-5. Visualization of various Transformer backbones Encoder-Only Architectures Encoder-only architectures were all the rage when Transformer-based language models first burst on the scene. Iconic language models from yesteryear (circa 2018) that use encoder-only architectures include BERT, RoBERTa, etc. There haven’t been many encoder-only LLMs trained

    参考来源

    - 位置:《Designing Large Language Model Applications》第04章《架构与学习目标》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Designing Large Language Model Applications/_epub-src/OEBPS/ch04.html) - 结论依据:摘录中关于编码器与解码器按可见性分工的定义、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:Encoder-only Encoder-decoder Decoder-only Let’s look at each of these in detail. Figure 4-5 depicts encoder-only, encoder-decoder, and decoder-only architectures. Figure 4-5. Visualization of various Transformer backbones Encoder-Only Architectures Encoder-only architectures were all the rage when Transformer-based language models first burst on the scene. Iconic language models from yesteryear (circa 2018) that use encoder-only architectures include BERT, RoBERTa, etc. There haven’t been many encoder-only LLMs trained