Enregistré dans:
Détails bibliographiques
Auteurs principaux: Ribeiro, Lucas Grativol, Leonardon, Mathieu, Muller, Guillaume, Fresse, Virginie, Arzel, Matthieu
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:https://arxiv.org/abs/2406.14082
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866911933978902528
author Ribeiro, Lucas Grativol
Leonardon, Mathieu
Muller, Guillaume
Fresse, Virginie
Arzel, Matthieu
author_facet Ribeiro, Lucas Grativol
Leonardon, Mathieu
Muller, Guillaume
Fresse, Virginie
Arzel, Matthieu
contents Low-Rank Adaptation (LoRA) methods have gained popularity in efficient parameter fine-tuning of models containing hundreds of billions of parameters. In this work, instead, we demonstrate the application of LoRA methods to train small-vision models in Federated Learning (FL) from scratch. We first propose an aggregation-agnostic method to integrate LoRA within FL, named FLoCoRA, showing that the method is capable of reducing communication costs by 4.8 times, while having less than 1% accuracy degradation, for a CIFAR-10 classification task with a ResNet-8. Next, we show that the same method can be extended with an affine quantization scheme, dividing the communication cost by 18.6 times, while comparing it with the standard method, with still less than 1% of accuracy loss, tested with on a ResNet-18 model. Our formulation represents a strong baseline for message size reduction, even when compared to conventional model compression works, while also reducing the training memory requirements due to the low-rank adaptation.
format Preprint
id arxiv_https___arxiv_org_abs_2406_14082
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle FLoCoRA: Federated learning compression with low-rank adaptation
Ribeiro, Lucas Grativol
Leonardon, Mathieu
Muller, Guillaume
Fresse, Virginie
Arzel, Matthieu
Machine Learning
Signal Processing
Low-Rank Adaptation (LoRA) methods have gained popularity in efficient parameter fine-tuning of models containing hundreds of billions of parameters. In this work, instead, we demonstrate the application of LoRA methods to train small-vision models in Federated Learning (FL) from scratch. We first propose an aggregation-agnostic method to integrate LoRA within FL, named FLoCoRA, showing that the method is capable of reducing communication costs by 4.8 times, while having less than 1% accuracy degradation, for a CIFAR-10 classification task with a ResNet-8. Next, we show that the same method can be extended with an affine quantization scheme, dividing the communication cost by 18.6 times, while comparing it with the standard method, with still less than 1% of accuracy loss, tested with on a ResNet-18 model. Our formulation represents a strong baseline for message size reduction, even when compared to conventional model compression works, while also reducing the training memory requirements due to the low-rank adaptation.
title FLoCoRA: Federated learning compression with low-rank adaptation
topic Machine Learning
Signal Processing
url https://arxiv.org/abs/2406.14082