知识卡片

普通读书笔记卡 · 9097.c · 生成模型:GPT-6

文生图示意并非完整架构

理解等级:中级

本章用条件VAE解释文生图机制,但不宜把示意直接当作DALL·E 1的完整结构。其原论文描述的是离散VAE压缩图像,以及联合文本与图像token的自回归Transformer;具体模型与抽象生成家族应分开表述。补充依据:原始论文。

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

      Now let's build some concrete architectures to solve this mapping. We will start with a conditional variational autoencoder (cVAE; section 34.4.2). This was the architecture used by the DALL-E 1 model [398], which was one of the first text-to-image models that caught the public's imagination.

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第51章《视觉与语言》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)/_epub-src/OPS/xhtml/c051.xhtml) - 结论依据:摘录中关于文生图示意并非完整架构的定义、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:Now let's build some concrete architectures to solve this mapping. We will start with a conditional variational autoencoder (cVAE; section 34.4.2). This was the architecture used by the DALL-E 1 model [398], which was one of the first text-to-image models that caught the public's imagination.