paper-with-me

홈 › Papers

MELoRA: Mini-Ensemble Low-Rank Adapters for Parameter-Efficient Fine-Tuning

2024-02-27 · Pengjie Ren, Chengshun Shi, Shiguang Wu, Mengqi Zhang, Zhaochun Ren, Maarten de Rijke, Zhumin Chen, Jiahuan Pei

Parameter-efficient fine-tuning (PEFT) is a popular method for tailoring pre-trained large language models (LLMs), especially as the models' scale and the diversity of tasks increase. Low-rank adaptation (LoRA) is based on the idea that the adaptation process is intrinsically low-dimensional, i.e., significant model changes can be represented with relatively few parameters. However, decreasing the rank encounters challenges with generalization errors for specific tasks when compared to full-parameter fine-tuning. We present MELoRA, a mini-ensemble low-rank adapters that uses fewer trainable parameters while maintaining a higher rank, thereby offering improved performance potential. The core idea is to freeze original pretrained weights and train a group of mini LoRAs with only a small number of parameters. This can capture a significant degree of diversity among mini LoRAs, thus promoting better generalization ability. We conduct a theoretical analysis and empirical studies on various NLP tasks. Our experimental results show that, compared to LoRA, MELoRA achieves better performance with 8 times fewer trainable parameters on natural language understanding tasks and 36 times fewer trainable parameters on instruction following tasks, which demonstrates the effectiveness of MELoRA.

📄 PDF Abstract BibTeX arXiv:2402.17263

Code (1)

chasonshi/melora 공식 구현 pytorch

Tasks

DiversityInstruction FollowingNatural Language Understandingparameter-efficient fine-tuning

Similar Papers 제목 키워드 기반

EulerLoRA: Rank-Driven Jump Dynamics for Calibrated Parameter-Efficient Fine-Tuning

2026-08-02 · Srinivas Anumasa, Dianbo Liu arxiv

Low-Rank Adaptation (LoRA) enables parameter-efficient fine-tuning, but standard LoRA produces a single deterministic model and does not directly support predictive uncertainty estimation. We introduce EulerLoRA, a stoch…

parameter-efficient fine-tuningOut-of-Distribution Detection

LoRA ensembles for large language model fine-tuning

2023-09-29 · Xi Wang, Laurence Aitchison, Maja Rudolph

Finetuned LLMs often exhibit poor uncertainty quantification, manifesting as overconfidence, poor calibration, and unreliable prediction results on test data or out-of-distribution samples. One approach commonly used in …

Language ModelingLanguage ModellingLarge Language Modelmodel+2

Efficient Test Time Adapter Ensembling for Low-resource Language Varieties

2021-09-10 · Findings (EMNLP) 2021 11 · Xinyi Wang, Yulia Tsvetkov, Sebastian Ruder, Graham Neubig

Adapters are light-weight modules that allow parameter-efficient fine-tuning of pretrained models. Specialized language and task adapters have recently been proposed to facilitate cross-lingual transfer of multilingual p…

Cross-Lingual Transfernamed-entity-recognitionNamed Entity RecognitionNamed Entity Recognition (NER)+3

BoostLoRA: Growing Effective Rank by Boosting Adapters

2026-04-30 · Raviteja Anantha, Nick Levato, Layne C. Price arxiv

Parameter-efficient fine-tuning (PEFT) methods face a tradeoff between adapter size and expressivity: ultra-low-parameter adapters are confined to fixed low-rank subspaces, capping performance even with extended training…

parameter-efficient fine-tuningCode Generation

A Little Rank Goes a Long Way: Random Scaffolds with LoRA Adapters Are All You Need

2026-04-09 · Hananel Hazan, Yanbo Zhang, Benedikt Hartl, Michael Levin arxiv

How many of a neural network's parameters actually encode task-specific information? We investigate this question with LottaLoRA, a training paradigm in which every backbone weight is drawn at random and frozen; only low…