Pruning All-Rounder: Rethinking and Improving Inference Efficiency for Large Vision Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Suo, Wei, Ma, Ji, Sun, Mengyang, Wu, Lin Yuanbo, Wang, Peng, Zhang, Yanning
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916871843872768
author Suo, Wei
Ma, Ji
Sun, Mengyang
Wu, Lin Yuanbo
Wang, Peng
Zhang, Yanning
author_facet Suo, Wei
Ma, Ji
Sun, Mengyang
Wu, Lin Yuanbo
Wang, Peng
Zhang, Yanning
contents Although Large Vision-Language Models (LVLMs) have achieved impressive results, their high computational costs pose a significant barrier to wide application. To enhance inference efficiency, most existing approaches can be categorized as parameter-dependent or token-dependent strategies to reduce computational demands. However, parameter-dependent methods require retraining LVLMs to recover performance while token-dependent strategies struggle to consistently select the most relevant tokens. In this paper, we systematically analyze the above challenges and provide a series of valuable insights for inference acceleration. Based on these findings, we propose a novel framework, the Pruning All-Rounder (PAR). Different from previous works, PAR develops a meta-router to adaptively organize pruning flows across both tokens and layers. With a self-supervised learning manner, our method achieves a superior balance between performance and efficiency. Notably, PAR is highly flexible, offering multiple pruning versions to address a range of acceleration scenarios. The code for this work is publicly available at https://github.com/ASGO-MM/Pruning-All-Rounder.
format Preprint
id arxiv_https___arxiv_org_abs_2412_06458
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Pruning All-Rounder: Rethinking and Improving Inference Efficiency for Large Vision Language Models
Suo, Wei
Ma, Ji
Sun, Mengyang
Wu, Lin Yuanbo
Wang, Peng
Zhang, Yanning
Computer Vision and Pattern Recognition
Although Large Vision-Language Models (LVLMs) have achieved impressive results, their high computational costs pose a significant barrier to wide application. To enhance inference efficiency, most existing approaches can be categorized as parameter-dependent or token-dependent strategies to reduce computational demands. However, parameter-dependent methods require retraining LVLMs to recover performance while token-dependent strategies struggle to consistently select the most relevant tokens. In this paper, we systematically analyze the above challenges and provide a series of valuable insights for inference acceleration. Based on these findings, we propose a novel framework, the Pruning All-Rounder (PAR). Different from previous works, PAR develops a meta-router to adaptively organize pruning flows across both tokens and layers. With a self-supervised learning manner, our method achieves a superior balance between performance and efficiency. Notably, PAR is highly flexible, offering multiple pruning versions to address a range of acceleration scenarios. The code for this work is publicly available at https://github.com/ASGO-MM/Pruning-All-Rounder.
title Pruning All-Rounder: Rethinking and Improving Inference Efficiency for Large Vision Language Models
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2412.06458