TLoRA: Task-aware Low Rank Adaptation of Large Language Models

TL;DR

TLoRA introduces a unified, data-driven approach to optimize initialization and resource allocation in low-rank adaptation of large language models, enhancing efficiency and performance across diverse NLP tasks.

Contribution

It proposes a novel framework that jointly optimizes initialization and rank allocation, improving parameter efficiency and task adaptability in LoRA-based fine-tuning.

Findings

TLoRA achieves superior performance across multiple NLP tasks.

It significantly reduces the number of trainable parameters.

The method demonstrates consistent improvements over existing LoRA variants.

Abstract

Low-Rank Adaptation (LoRA) has become a widely adopted parameter-efficient fine-tuning method for large language models, with its effectiveness largely influenced by the allocation of ranks and scaling factors, as well as initialization. Existing LoRA variants typically address only one of these factors, often at the cost of increased training complexity or reduced practical efficiency. In this work, we present Task-aware Low-Rank Adaptation (TLoRA), a unified framework that jointly optimizes initialization and resource allocation at the outset of training. TLoRA introduces a data-driven initialization strategy that aligns the LoRA $A$ matrix with task-relevant subspaces by performing singular value decomposition on the product of pre-trained weights and input activation covariance. After this, the $A$ matrix is frozen, and only the $B$ matrix is trained. Furthermore, TLoRA employs a sensitivity-based importance metric to adaptively allocate ranks and scaling factors across layers under a fixed parameter budget. We conduct extensive experiments that demonstrate TLoRA consistently performs excellently across various tasks, including natural language understanding, commonsense reasoning, math reasoning, code generation, and chat generation, while significantly reducing the number of trainable parameters.