Rethinking Pruning for Vision-Language Models: Strategies for Effective Sparsity and Performance Restoration
This work addresses the problem of model compression for Vision-Language Models, which is incremental as it builds on existing pruning and finetuning techniques.
The study tackled the challenge of deploying large-scale Vision-Language Models in resource-constrained scenarios by developing pruning strategies and a method called SparseLoRA to restore performance, achieving improvements such as an 11.3% boost under 2:4 sparsity and a 47.6% enhancement under unstructured 70% sparsity.
Vision-Language Models (VLMs) integrate information from multiple modalities and have shown remarkable success across various tasks. However, deploying large-scale VLMs in resource-constrained scenarios is challenging. Pruning followed by finetuning offers a potential solution but remains underexplored for VLMs. This study addresses two key questions: how to distribute sparsity across different modality-specific models, and how to restore the performance of pruned sparse VLMs. Our preliminary studies identified two effective pruning settings: applying the same sparsity to both vision and language models, and pruning only the language models. While LoRA finetuning aims to restore sparse models, it faces challenges due to incompatibility with sparse models, disrupting the pruned sparsity. To overcome these issues, we propose SparseLoRA, which applies sparsity directly to LoRA weights. Our experimental results demonstrate significant improvements, including an 11.3\% boost under 2:4 sparsity and a 47.6\% enhancement under unstructured 70\% sparsity. Code is released at: \url{https://github.com/Shwai-He/VLM-Compression}.