LoRA-drop: Efficient LoRA Parameter Pruning based on Output Evaluation

TL;DR

LoRA-drop is a novel method that prunes LoRA parameters based on their output importance, reducing resource consumption while maintaining performance in fine-tuning large models.

Contribution

It introduces a new importance evaluation based on LoRA output, enabling effective parameter pruning during fine-tuning.

Findings

Achieves comparable performance to full fine-tuning and LoRA.

Retains 50% of LoRA parameters on average.

Effective across various model scales and NLP tasks.

Abstract

Low-Rank Adaptation (LoRA) is currently the most commonly used Parameter-efficient fine-tuning (PEFT) method, it introduces auxiliary parameters for each layer to fine-tune the pre-trained model under limited computing resources. However, it still faces resource consumption challenges during training when scaling up to larger models. Most previous studies have tackled this issue by using pruning techniques, which involve removing LoRA parameters deemed unimportant. Nonetheless, these efforts only analyze LoRA parameter features to evaluate their importance, such as parameter count, size, and gradient. In fact, the output of LoRA (product of LoRA parameter and hidden state), directly impacts the final results. Preliminary experiments indicate that a fraction of LoRA elements possesses significantly high output values, substantially influencing the layer output. Motivated by the observation, we propose LoRA-drop. Concretely, LoRA-drop evaluates the importance of LoRA based on the LoRA output. Then we retain LoRA for important layers and the other layers share the same LoRA. We conduct abundant experiments with models of different scales on NLU and NLG tasks. Results demonstrate that LoRA-drop can achieve performance comparable to full fine-tuning and LoRA, while retaining 50\% of the LoRA parameters on average.