Pruning as a Domain-specific LLM Extractor

TL;DR

This paper presents D-Pruner, a novel unstructured dual-pruning method that efficiently compresses large language models for domain-specific applications by preserving both general capabilities and domain knowledge.

Contribution

It introduces a new pruning technique that identifies and retains crucial weights for both general language skills and domain-specific knowledge, improving domain adaptation of LLMs.

Findings

Effective domain-specific compression demonstrated in healthcare and legal tasks.

Preserves general language capabilities while reducing model size.

Outperforms existing pruning methods in domain adaptation scenarios.

Abstract

Large Language Models (LLMs) have exhibited remarkable proficiency across a wide array of NLP tasks. However, the escalation in model size also engenders substantial deployment costs. While few efforts have explored model pruning techniques to reduce the size of LLMs, they mainly center on general or task-specific weights. This leads to suboptimal performance due to lacking specificity on the target domain or generality on different tasks when applied to domain-specific challenges. This work introduces an innovative unstructured dual-pruning methodology, D-Pruner, for domain-specific compression on LLM. It extracts a compressed, domain-specific, and task-agnostic LLM by identifying LLM weights that are pivotal for general capabilities, like linguistic capability and multi-task solving, and domain-specific knowledge. More specifically, we first assess general weight importance by quantifying the error incurred upon their removal with the help of an open-domain calibration dataset. Then, we utilize this general weight importance to refine the training loss, so that it preserves generality when fitting into a specific domain. Moreover, by efficiently approximating weight importance with the refined training loss on a domain-specific calibration dataset, we obtain a pruned model emphasizing generality and specificity. Our comprehensive experiments across various tasks in healthcare and legal domains show the effectiveness of D-Pruner in domain-specific compression. Our code is available at https://github.com/psunlpgroup/D-Pruner.