PRILoRA: Pruned and Rank-Increasing Low-Rank Adaptation

TL;DR

PRILoRA enhances parameter-efficient fine-tuning of large language models by adaptively allocating ranks and pruning layers, leading to improved performance on multiple benchmarks.

Contribution

It introduces a novel method that dynamically allocates layer-specific ranks and prunes during training, improving upon existing low-rank adaptation techniques.

Findings

Sets new state-of-the-art results on eight GLUE benchmarks.

Effectively allocates ranks across layers for better model adaptation.

Demonstrates significant efficiency improvements in fine-tuning large models.

Abstract

With the proliferation of large pre-trained language models (PLMs), fine-tuning all model parameters becomes increasingly inefficient, particularly when dealing with numerous downstream tasks that entail substantial training and storage costs. Several approaches aimed at achieving parameter-efficient fine-tuning (PEFT) have been proposed. Among them, Low-Rank Adaptation (LoRA) stands out as an archetypal method, incorporating trainable rank decomposition matrices into each target module. Nevertheless, LoRA does not consider the varying importance of each layer. To address these challenges, we introduce PRILoRA, which linearly allocates a different rank for each layer, in an increasing manner, and performs pruning throughout the training process, considering both the temporary magnitude of weights and the accumulated statistics of the input to any given layer. We validate the effectiveness of PRILoRA through extensive experiments on eight GLUE benchmarks, setting a new state of the art.