부산대학교 도서관

Structured pruning of modern large language models (LLMs) has emerged as a way of decreasing their high computational needs. Width pruning reduces the size of projection weight matrices (e.g., by removing attention heads) while maintaining the number of layers. Depth pruning, in contrast, removes entire layers or blocks, while keeping the size of the remaining weights unchanged. Most current research focuses on either width-only or a blend of width and depth pruning, with little comparative analysis between the two units (width vs. depth) concerning their impact on LLM inference efficiency. In this work, we show that simple depth pruning can effectively compress LLMs while achieving comparable or superior performance to recent width pruning studies. Our pruning method boosts inference speeds, especially under memory-constrained conditions that require limited batch sizes for running LLMs, where width pruning is ineffective. In retraining pruned models for quality recovery, continued pretraining on a large corpus markedly outperforms LoRA-based tuning, particularly at severe pruning ratios. We hope this work can help build compact yet capable LLMs. Code and models can be found at: https://github.com/Nota-NetsPresso/shortened-llm
Comment: Update (arXiv-v2): continued pretraining for severe pruning ratios, compatibility with quantization, and enhanced baselines. Preliminary work (arXiv-v1) accepted at ICLR 2024 Workshop on ME-FoMo: https://openreview.net/forum?id=18VGxuOdpu