NutePrune: Efficient Progressive Pruning with Numerous Teachers for Large Language Models

TL;DR

NutePrune is a resource-efficient progressive pruning method for large language models that uses multiple teachers and LoRA modules to achieve high performance at significant sparsity levels.

Contribution

It introduces an efficient multi-teacher pruning approach that reduces memory costs and improves pruning effectiveness for large language models.

Findings

Retains 97.17% of original performance at 20% sparsity

Achieves 95.07% performance at 25% sparsity

Demonstrates effectiveness across various tasks

Abstract

The considerable size of Large Language Models (LLMs) presents notable deployment challenges, particularly on resource-constrained hardware. Structured pruning, offers an effective means to compress LLMs, thereby reducing storage costs and enhancing inference speed for more efficient utilization. In this work, we study data-efficient and resource-efficient structure pruning methods to obtain smaller yet still powerful models. Knowledge Distillation is well-suited for pruning, as the intact model can serve as an excellent teacher for pruned students. However, it becomes challenging in the context of LLMs due to memory constraints. To address this, we propose an efficient progressive Numerous-teacher pruning method (NutePrune). NutePrune mitigates excessive memory costs by loading only one intact model and integrating it with various masks and LoRA modules, enabling it to seamlessly switch between teacher and student roles. This approach allows us to leverage numerous teachers with varying capacities to progressively guide the pruned model, enhancing overall performance. Extensive experiments across various tasks demonstrate the effectiveness of NutePrune. In LLaMA-7B zero-shot experiments, NutePrune retains 97.17% of the performance of the original model at 20% sparsity and 95.07% at 25% sparsity. Our code is available at https://github.com/Lucius-lsr/NutePrune.