Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

paper CoOP #474

Open
junxnone opened this issue Oct 8, 2024 · 0 comments
Open

paper CoOP #474

junxnone opened this issue Oct 8, 2024 · 0 comments

Comments

@junxnone
Copy link
Owner

junxnone commented Oct 8, 2024

CoOP

  • Context Optimization(CoOP) - 通过自动优化提示词,提高 VLM 模型下游任务的性能,减少人工提示工程的工作量
  • PET Prompt --> Best pizza ever! It was ___
  • AutoPrompt --> [T] [T] [T] [T] [T] [P]
  • CLIP --> A photo of a {CLASS} :: [Image]
  • CoOP --> [V] [V] [V] [V] [V] [CLASS] :: [Image]
  • AnomlyCLIP/FILo 等异常检测使用了 CoOP 方法

image

CoCoOP

  • 解决 CoOp 方法中学习到的上下文无法推广到同一数据集中未见过的类别的问题。
  • 条件提示学习,使提示能够根据每个输入实例进行调整,提高模型的泛化能力
  • 在 CoOp 的基础上,进一步学习一个轻量级神经网络,为每个图像生成一个输入条件令牌(向量),与可学习的上下文向量相结合。

image

CoCoOP vs CoOP

image

Reference

@junxnone junxnone changed the title x CoOP paper CoOP Oct 16, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant