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.

Sign up for GitHub

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

Jump to bottom

[27] AdaptFormer: Adapting Vision Transformers for Scalable Visual Recognition #27

Open

Dongwoo-Im opened this issue Apr 20, 2023 · 1 comment

Assignees

Labels

Efficient Finetuning NIPS '22 Visual Adapter

Contributor

Dongwoo-Im commented Apr 20, 2023

Links

Paper : https://arxiv.org/abs/2205.13535
Github : https://github.com/ShoufaChen/AdaptFormer

한 줄 요약

ViT의 FFN module에 AdaptMLP branch를 residual connection하는 AdaptFormer 제안
Image, Video classification 에서 성능을 검증하였지만, dense prediction에 대한 연구가 포함되지 않은 것은 아쉽다.
아직 읽어보진 않았지만 관심 있다면 Visual Prompt Tuning (VPT), ViT-Adapter 논문도 읽어보면 좋을 것 같다.

선택 이유

Adapter 모듈이 network 구조에 어떻게 적용될 수 있는지 궁금하여 찾아보다가 선택

Dongwoo-Im added NIPS '22 Efficient Finetuning labels

Dongwoo-Im self-assigned this

Contributor Author

Dongwoo-Im commented Apr 20, 2023

https://dongwoo-im.github.io/papers/review/2023-04-19-AdaptFormer/

Dongwoo-Im added the Visual Adapter label

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment