Crossmark

Structural Pruning of Large Vision Language Models: A Comprehensive Study on Pruning Dynamics, Recovery, and Data Efficiency

Published Online: 2026-06-09

Published Print: 2026-06

Authors

Huang, Yiran https://orcid.org/0009-0004-6937-5668
Thede, Lukas

Mancini, Massimiliano

Xu, Wenjia

Akata, Zeynep
Funding

Funding for this research was provided by:

Technische Universität München
License Information

Text and Data Mining valid from 2026-06-01

Version of Record valid from 2026-06-09
More Information

Article History

Received: 15 January 2026

Accepted: 24 April 2026

First Online: 9 June 2026

Document is current