Skip to main navigation Skip to search Skip to main content

SmartTrim: Adaptive Tokens and Attention Pruning for Efficient Vision-Language Models

  • Zekun Wang*
  • , Jingchang Chen*
  • , Wangchunshu Zhou
  • , Haichao Zhu Jiafeng Liang
  • , Liping Shan
  • , Ming Liu
  • , Dongliang Xu
  • , Qing Yang
  • , Bing Qin
  • *Corresponding author for this work
  • Harbin Institute of Technology
  • ETH Zurich
  • Du Xiaoman (Beijing) Science Technology Co.,Ltd
  • Peng Cheng Laboratory

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Despite achieving remarkable performance on various vision-language tasks, Transformer-based Vision-Language Models (VLMs) suffer from redundancy in inputs and parameters, significantly hampering their efficiency in real-world applications. Moreover, the degree of redundancy in token representations and model parameters, such as attention heads, varies significantly for different inputs. In light of the challenges, we propose SmartTrim, an adaptive acceleration framework for VLMs, which adjusts the computational overhead per instance. Specifically, we integrate lightweight modules into the original backbone to identify and prune redundant token representations and attention heads within each layer. Furthermore, we devise a self-distillation strategy to enhance the consistency between the predictions of the pruned model and its fully-capacity counterpart. Experimental results across various vision-language tasks consistently demonstrate that SmartTrim accelerates the original model by 2-3 times with minimal performance degradation, highlighting the effectiveness and efficiency compared to previous approaches. Code will be available at https://github.com/kugwzk/SmartTrim.

Original languageEnglish
Title of host publication2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation, LREC-COLING 2024 - Main Conference Proceedings
EditorsNicoletta Calzolari, Min-Yen Kan, Veronique Hoste, Alessandro Lenci, Sakriani Sakti, Nianwen Xue
PublisherEuropean Language Resources Association (ELRA)
Pages14937-14953
Number of pages17
ISBN (Electronic)9782493814104
StatePublished - 2024
EventJoint 30th International Conference on Computational Linguistics and 14th International Conference on Language Resources and Evaluation, LREC-COLING 2024 - Hybrid, Torino, Italy
Duration: 20 May 202425 May 2024

Publication series

Name2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation, LREC-COLING 2024 - Main Conference Proceedings

Conference

ConferenceJoint 30th International Conference on Computational Linguistics and 14th International Conference on Language Resources and Evaluation, LREC-COLING 2024
Country/TerritoryItaly
CityHybrid, Torino
Period20/05/2425/05/24

Keywords

  • Adaptive Inference
  • Dynamic Network
  • Pruning
  • Vision-Language Model

Fingerprint

Dive into the research topics of 'SmartTrim: Adaptive Tokens and Attention Pruning for Efficient Vision-Language Models'. Together they form a unique fingerprint.

Cite this