Visual Prompt Tuning
Publikation: Bidrag til bog/antologi/rapport › Konferencebidrag i proceedings › Forskning › fagfællebedømt
The current modus operandi in adapting pre-trained models involves updating all the backbone parameters, i.e., full fine-tuning. This paper introduces Visual Prompt Tuning (VPT) as an efficient and effective alternative to full fine-tuning for large-scale Transformer models in vision. Taking inspiration from recent advances in efficiently tuning large language models, VPT introduces only a small amount (less than 1% of model parameters) of trainable parameters in the input space while keeping the model backbone frozen. Via extensive experiments on a wide variety of downstream recognition tasks, we show that VPT achieves significant performance gains compared to other parameter efficient tuning protocols. Most importantly, VPT even outperforms full fine-tuning in many cases across model capacities and training data scales, while reducing per-task storage cost. Code is available at github.com/kmnp/vpt.
Originalsprog | Engelsk |
---|---|
Titel | Computer Vision – ECCV 2022 : 17th European Conference, Proceedings |
Redaktører | Shai Avidan, Gabriel Brostow, Moustapha Cissé, Giovanni Maria Farinella, Tal Hassner |
Antal sider | 19 |
Forlag | Springer |
Publikationsdato | 2022 |
Sider | 709-727 |
ISBN (Trykt) | 978-3-031-19826-7 |
ISBN (Elektronisk) | 978-3-031-19827-4 |
DOI | |
Status | Udgivet - 2022 |
Begivenhed | 17th European Conference on Computer Vision, ECCV 2022 - Tel Aviv, Israel Varighed: 23 okt. 2022 → 27 okt. 2022 |
Konference
Konference | 17th European Conference on Computer Vision, ECCV 2022 |
---|---|
Land | Israel |
By | Tel Aviv |
Periode | 23/10/2022 → 27/10/2022 |
Navn | Lecture Notes in Computer Science |
---|---|
Vol/bind | 13693 LNCS |
ISSN | 0302-9743 |
Bibliografisk note
Funding Information:
Acknowledgement. Menglin is supported by a Meta AI research grant awarded to Cornell University, Luming and Bharath is supported by NSF IIS-2144117, Serge is supported in part by the Pioneer Centre for AI, DNRF grant number P1. We would like to thank Alexander Rush, Yin Cui for valuable suggestions and discussion.
Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
ID: 342671827