SSMLoRA: Enhancing Low-Rank Adaptation with State Space Model

TL;DR

SSMLoRA enhances low-rank adaptation for language models by integrating state space models, maintaining performance with fewer parameters and better handling longer sequences, compared to traditional LoRA methods.

Contribution

Introduces SSMLoRA, a novel extension of LoRA that incorporates state space models to improve parameter efficiency and performance in low-rank adaptation.

Findings

Achieves comparable GLUE benchmark performance with half the parameters.

Maintains performance with sparser insertions.

Shows potential for longer input sequence tasks.

Abstract

Fine-tuning is a key approach for adapting language models to specific downstream tasks, but updating all model parameters becomes impractical as model sizes increase. Parameter-Efficient Fine-Tuning (PEFT) methods, such as Low-Rank Adaptation (LoRA), address this challenge by introducing additional adaptation parameters into pre-trained weight matrices. However, LoRA's performance varies across different insertion points within the model, highlighting potential parameter inefficiency due to unnecessary insertions. To this end, we propose SSMLoRA (State Space Model Low-Rank Adaptation), an extension of LoRA that incorporates a State Space Model (SSM) to interconnect low-rank matrices. SSMLoRA ensures that performance is maintained even with sparser insertions. SSMLoRA allows the model to not only map inputs to a low-rank space for better feature extraction but also leverage the computations from the previous low-rank space. Our method achieves comparable performance to LoRA on the General Language Understanding Evaluation (GLUE) benchmark while using only half the parameters. Additionally, due to its structure, SSMLoRA shows promise in handling tasks with longer input sequences. .You can find our code here:https://github.com/yuhkalhic/SSMLoRA.