ScaLoRA: Optimally Scaled Low-Rank Adaptation for Efficient High-Rank Fine-Tuning

TL;DR

ScaLoRA introduces a method to enhance low-rank adaptation for large language models by optimally scaling incremental updates, leading to improved efficiency and performance in fine-tuning tasks.

Contribution

The paper proposes a novel approach that accumulates high-rank weight updates through optimal scaling of low-rank matrices, overcoming limitations of traditional LoRA.

Findings

Consistent performance improvements over state-of-the-art LoRA variants.

Faster convergence in fine-tuning large language models.

Effective across diverse NLP tasks including reasoning and problem solving.

Abstract

As large language models (LLMs) continue to scale in size, the computational overhead has become a major bottleneck for task-specific fine-tuning. While low-rank adaptation (LoRA) effectively curtails this cost by confining the weight updates to a low-dimensional subspace, such a restriction can hinder effectiveness and slow convergence. This contribution deals with these limitations by accumulating progressively a high-rank weight update from consecutive low-rank increments. Specifically, the per update optimal low-rank matrix is identified to minimize the loss function and closely approximate full fine-tuning. To endow efficient and seamless optimization without restarting, this optimal choice is formed by appropriately scaling the columns of the original low-rank matrix. Rigorous performance guarantees reveal that the optimal scaling can be found analytically. Extensive numerical tests with popular LLMs scaling up to 12 billion parameters demonstrate a consistent performance gain and fast convergence relative to state-of-the-art LoRA variants on diverse tasks including natural language understanding, commonsense reasoning, and mathematical problem solving.