You Need Multiple Exiting: Dynamic Early Exiting for Accelerating Unified Vision Language Model-Reference-Cited by-同舟云学术

You Need Multiple Exiting: Dynamic Early Exiting for Accelerating Unified Vision Language Model

Published:2023-06 Issue: Volume: Page:
ISSN:
Container-title:2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
language:
Short-container-title:

Author:

Tang Shengkun¹,Wang Yaqing²,Kong Zhenglun³,Zhang Tianchi⁴,Li Yao⁵,Ding Caiwen⁶,Wang Yanzhi³,Liang Yi²,Xu Dongkuan¹

Affiliation:

1. North Carolina State University,Raleigh,USA

2. Google Research,New York,USA

3. Northeastern University,Boston,USA

4. University of Michigan,Ann Arbor,USA

5. The University of North Carolina at Chapel Hill,Chapel Hill,USA

6. University of Connecticut,Mansfield,USA

Publisher

IEEE

Link

Reference60 articles.

1. Distilling the knowledge in a neural network;hinton;ArXiv Preprint,2015

4. Coca: Contrastive captioners are image-text foundation models;jiahui;ArXiv Preprint,2022

5. Shallow-deep networks: Understanding and mitigating network overthinking;kaya;In International Conference on Machine Learning,2019

Cited by 4 articles. 订阅此论文施引文献订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献

3. Efficient Controllable Multi-Task Architectures;2023 IEEE/CVF International Conference on Computer Vision (ICCV);2023-10-01

4. Efficient Image-Text Retrieval via Keyword-Guided Pre-Screening;IEEE Transactions on Circuits and Systems for Video Technology;2023