SPAP: Structured Pruning via Alternating Optimization and Penalty Methods
This work addresses the deployment constraints of large language models for users needing efficient inference, offering a practical solution with incremental improvements over existing pruning methods.
The paper tackles the problem of high computational and memory demands in large language models by proposing SPAP, a structured pruning framework based on optimization theory, which achieves linear inference speedups of 1.29x at 30% sparsity and proportional memory reductions while preserving performance.
The deployment of large language models (LLMs) is often constrained by their substantial computational and memory demands. While structured pruning presents a viable approach by eliminating entire network components, existing methods suffer from performance degradation, reliance on heuristic metrics, or expensive finetuning. To address these challenges, we propose SPAP (Structured Pruning via Alternating Optimization and Penalty Methods), a novel and efficient structured pruning framework for LLMs grounded in optimization theory. SPAP formulates the pruning problem through a mixed-integer optimization model, employs a penalty method that effectively makes pruning decisions to minimize pruning errors, and introduces an alternating minimization algorithm tailored to the splittable problem structure for efficient weight updates and performance recovery. Extensive experiments on OPT, LLaMA-3/3.1/3.2, and Qwen2.5 models demonstrate SPAP's superiority over state-of-the-art methods, delivering linear inference speedups (1.29$\times$ at 30% sparsity) and proportional memory reductions. Our work offers a practical, optimization-driven solution for pruning LLMs while preserving model performance.