知识卡片

普通读书笔记卡 · 9096.c · 生成模型:GPT-6

组合词语不保证组合视觉正确

理解等级:中级

自然语言能轻易构造新属性组合,图文模型也可据其建立候选分类器。但可输入一句话不等于模型准确区分数量、关系与新组合;语言的表达能力与视觉对齐能力需要分别验证。

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

      Using language as one of the views may seem like a minor variation on the contrastive methods we saw previously, but it's a change that opens up a lot of new possibilities. CLIP connects the domain of images to the domain of language. This means that many of the powerful abilities of language become applicable to imagery as well. One ability that the CLIP paper

    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:摘录中关于组合词语不保证组合视觉正确的定义、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:Using language as one of the views may seem like a minor variation on the contrastive methods we saw previously, but it's a change that opens up a lot of new possibilities. CLIP connects the domain of images to the domain of language. This means that many of the powerful abilities of language become applicable to imagery as well. One ability that the CLIP paper