Compresso: Structured Pruning with Collaborative Prompting Learns Compact Large Language Models

TL;DR

Compresso introduces a collaborative prompt-based structured pruning method that efficiently compresses large language models like LLaMA-7B, maintaining or improving performance while significantly reducing model size.

Contribution

It presents a novel pruning paradigm combining resource-efficient algorithms and collaborative prompting, enabling effective structured pruning during instruction tuning.

Findings

Pruned LLaMA-7B to 5.4B with maintained or improved performance.

Outperformed one-shot pruning baselines across multiple benchmarks.

Achieved up to 11.43% higher scores in key NLP tasks.

Abstract

Despite the remarkable success of Large Language Models (LLMs), the massive size poses significant deployment challenges, particularly on resource-constrained hardware. While existing LLM compression methods focus on quantization, pruning remains relatively unexplored due to the high cost of training-based approaches and data collection challenges. One-shot pruning methods, although cost-effective and data-free, have become dominant in LLM pruning, but lead to performance decline under the structured pruning setting. In this work, we introduce a new paradigm for structurally pruning LLMs, called Compresso. Our approach, through the collaboration of the proposed resource-efficient pruning algorithm and the LLM itself, learns optimal pruning decisions during the training process. Compresso addresses the challenges of expensive training costs and data collection by incorporating Low-Rank Adaptation (LoRA) into the $L_0$ regularization during the instruction tuning process. Then, we further augment the pruning algorithm by introducing a collaborative prompt that fosters collaboration between the LLM and the pruning algorithm, significantly boosting the overall performance. To this end, Compresso prunes LLaMA-7B to 5.4B, maintaining original performance and even surpassing LLaMA-7B in reading comprehension by 2.62%. Extensive experiments demonstrate that Compresso significantly outperforms one-shot pruning baselines across various sparsity ratios, achieving up to 2.21%, 11.43%, 7.04%, and 4.81% higher scores on the commonsense reasoning, reading comprehension, MMLU, and BBH benchmarks, respectively.