知识卡片
可提示系统也需要训练
系统可以在训练时学习理解文字、笔画等人可解释条件,使用户在推理时通过这些输入调节行为。它与只优化无语义向量的提示不同;条件生成也是构建可提示系统的一种方式。给出输入条件和改参数是不同的适配通道。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 迁移学习与适配
原书出处 · 1 处
- · 查看原书依据
It is also possible to train a model to be promptable, so that at test time it adapts better to user prompts. An advantage of this approach is that we can train a model to understand human-interpretable prompts. For example, we could train the model so that its behavior can be adapted based on text instructions or based on user-drawn strokes (a representative example of
参考来源
- 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第37章《迁移学习与适配》(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/c037.xhtml)
- 结论依据:摘录中关于可提示系统也需要训练的定义、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:It is also possible to train a model to be promptable, so that at test time it adapts better to user prompts. An advantage of this approach is that we can train a model to understand human-interpretable prompts. For example, we could train the model so that its behavior can be adapted based on text instructions or based on user-drawn strokes (a representative example of