AutoPEFT: Automatic Configuration Search for Parameter-Efficient Fine-Tuning

TL;DR

AutoPEFT automates the search for optimal parameter-efficient fine-tuning configurations in large language models, achieving superior performance and efficiency trade-offs across multiple NLP tasks.

Contribution

It introduces an automated, multi-objective Bayesian optimization approach to discover effective PEFT configurations, outperforming manual designs and comparable to full fine-tuning.

Findings

AutoPEFT configurations outperform existing PEFT methods.

Discovered configurations are highly transferable across tasks.

AutoPEFT achieves performance comparable to full fine-tuning without high costs.

Abstract

Large pretrained language models are widely used in downstream NLP tasks via task-specific fine-tuning, but such procedures can be costly. Recently, Parameter-Efficient Fine-Tuning (PEFT) methods have achieved strong task performance while updating much fewer parameters than full model fine-tuning (FFT). However, it is non-trivial to make informed design choices on the PEFT configurations, such as their architecture, the number of tunable parameters, and even the layers in which the PEFT modules are inserted. Consequently, it is highly likely that the current, manually designed configurations are suboptimal in terms of their performance-efficiency trade-off. Inspired by advances in neural architecture search, we propose AutoPEFT for automatic PEFT configuration selection: we first design an expressive configuration search space with multiple representative PEFT modules as building blocks. Using multi-objective Bayesian optimisation in a low-cost setup, we then discover a Pareto-optimal set of configurations with strong performance-cost trade-offs across different numbers of parameters that are also highly transferable across different tasks. Empirically, on GLUE and SuperGLUE tasks, we show that AutoPEFT-discovered configurations significantly outperform existing PEFT methods and are on par or better than FFT without incurring substantial training efficiency costs.