LASER: Loss-Aware Singular-value Decomposition and Rank Allocation for Efficient Low-Precision Vision-Language Models
About
Vision-language models (VLMs) deliver strong multimodal reasoning capabilities, but their large computational cost and high parameter counts make deployment challenging on resource-constrained devices. Low-rank decomposition has emerged as a promising compression technique, yet existing methods often optimize local matrix reconstruction error, rely on uniform or heuristic rank allocation, and focus mainly on attention projections while leaving feed-forward networks underexplored. In this paper, we propose~\textit{LASER} (\textbf{L}oss-\textbf{A}ware \textbf{S}ingular-value d\textbf{E}composition and \textbf{R}ank allocation), a low-rank compression framework for efficient low-precision VLM inference. LASER derives a curvature-weighted SVD objective from a second-order approximation of the model loss and uses Kronecker-factored Fisher information to guide decomposition toward downstream performance rather than reconstruction alone. We further introduce a loss-aware cross-layer rank allocation strategy based on calibration gradients, enabling more effective parameter budgeting across layers. Finally, we extend low-rank compression to FFN layers through a hybrid scheme that combines SVD with quantization. The evaluation results show that LASER achieves more than $2.3\times$ decoding speedup over previous work while preserving strong accuracy under low-precision inference.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Multimodal Understanding | SEED-Bench | Accuracy71.4 | 516 | |
| Multimodal Science Question Answering | ScienceQA IMG | Accuracy72.11 | 152 | |
| Visual Question Answering | ScienceQA IMG | Accuracy84.58 | 135 | |
| Multimodal Reasoning | MMBench EN v1.1 | Accuracy80.68 | 125 | |
| Visual Question Answering | SciQA-IMG | Accuracy72.11 | 71 | |
| Multimodal Large Language Model Evaluation | SEED-Bench | Accuracy71.56 | 18 | |
| Multimodal Question Answering | SEED Bench Img | Accuracy71.56 | 18 | |
| Multimodal Question Answering | MMBench EN v1.1 | Accuracy65.08 | 18 | |
| Multimodal Reasoning | ScienceQA IMG (test) | Accuracy84.58 | 9 | |
| Multimodal Reasoning | Seed-Bench (test) | Accuracy76.22 | 7 |