ARMOR: High-Performance Semi-Structured Pruning via Adaptive Matrix Factorization
This work addresses deployment challenges for large language models by improving the trade-off between compression and accuracy in semi-structured pruning, offering a more effective solution for practical hardware acceleration.
The paper tackles the problem of performance degradation in semi-structured pruning for large language models by introducing ARMOR, a one-shot post-training pruning algorithm that factorizes weight matrices into a 2:4 sparse core with block diagonal wrappers, achieving superior performance over state-of-the-art methods on tasks like perplexity and downstream evaluations while maintaining inference speedups and memory reductions.
Large language models (LLMs) present significant deployment challenges due to their immense computational and memory requirements. While semi-structured pruning, particularly 2:4 sparsity, offers a path to practical hardware acceleration, existing methods often incur substantial performance degradation. To bridge this gap, we introduce ARMOR: (Adaptive Representation with Matrix-factORization), a novel one-shot post-training pruning algorithm. Instead of directly pruning weights, ARMOR factorizes each weight matrix into a 2:4 sparse core wrapped by two low-overhead, block diagonal matrices. These wrappers act as efficient pre and post-transformation error correctors, offering greater flexibility to preserve model quality compared to conventional 2:4 pruning techniques. The sparse core and block diagonal wrappers are chosen through a block coordinate descent algorithm that minimizes a layer-wise proxy loss. We theoretically prove this optimization is guaranteed to converge to a solution with a proxy loss less than or equal to state-of-the-art pruning algorithms. Experiments on Llama (Touvron et al., 2023; Dubey et al., 2024) and Qwen (Yang et al., 2025) model families demonstrate that ARMOR consistently and significantly outperforms state-of-the-art 2:4 pruning methods across a wide range of downstream tasks and perplexity evaluations. ARMOR achieves this superior performance while retaining the inference speedups and substantial memory usage reductions of 2:4 pruning, establishing a more effective trade-off between model compression and task accuracy