Enregistré dans:
Détails bibliographiques
Auteurs principaux: Savadikar, Chinmay, Song, Xi, Wu, Tianfu
Format: Preprint
Publié: 2023
Sujets:
Accès en ligne:https://arxiv.org/abs/2312.00700
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866912477515612160
author Savadikar, Chinmay
Song, Xi
Wu, Tianfu
author_facet Savadikar, Chinmay
Song, Xi
Wu, Tianfu
contents Fine-tuning large pretrained Transformer models can focus on either introducing a small number of new learnable parameters (parameter efficiency) or editing representations of a small number of tokens using lightweight modules (representation efficiency). While the pioneering method LoRA (Low-Rank Adaptation) inherently balances parameter, compute, and memory efficiency, many subsequent variants trade off compute and memory efficiency and/or performance to further reduce fine-tuning parameters. To address this limitation and unify parameter-efficient and representation-efficient fine-tuning, we propose Weight-Generative Fine-Tuning (WeGeFT, pronounced wee-gift), a novel approach that learns to generate fine-tuning weights directly from the pretrained weights. WeGeFT employs a simple low-rank formulation consisting of two linear layers, either shared across multiple layers of the pretrained model or individually learned for different layers. This design achieves multi-faceted efficiency in parameters, representations, compute, and memory, while maintaining or exceeding the performance of LoRA and its variants. Extensive experiments on commonsense reasoning, arithmetic reasoning, instruction following, code generation, and visual recognition verify the effectiveness of our proposed WeGeFT. Our code is available at https://github.com/savadikarc/wegeft
format Preprint
id arxiv_https___arxiv_org_abs_2312_00700
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle WeGeFT: Weight-Generative Fine-Tuning for Multi-Faceted Efficient Adaptation of Large Models
Savadikar, Chinmay
Song, Xi
Wu, Tianfu
Computer Vision and Pattern Recognition
Machine Learning
Fine-tuning large pretrained Transformer models can focus on either introducing a small number of new learnable parameters (parameter efficiency) or editing representations of a small number of tokens using lightweight modules (representation efficiency). While the pioneering method LoRA (Low-Rank Adaptation) inherently balances parameter, compute, and memory efficiency, many subsequent variants trade off compute and memory efficiency and/or performance to further reduce fine-tuning parameters. To address this limitation and unify parameter-efficient and representation-efficient fine-tuning, we propose Weight-Generative Fine-Tuning (WeGeFT, pronounced wee-gift), a novel approach that learns to generate fine-tuning weights directly from the pretrained weights. WeGeFT employs a simple low-rank formulation consisting of two linear layers, either shared across multiple layers of the pretrained model or individually learned for different layers. This design achieves multi-faceted efficiency in parameters, representations, compute, and memory, while maintaining or exceeding the performance of LoRA and its variants. Extensive experiments on commonsense reasoning, arithmetic reasoning, instruction following, code generation, and visual recognition verify the effectiveness of our proposed WeGeFT. Our code is available at https://github.com/savadikarc/wegeft
title WeGeFT: Weight-Generative Fine-Tuning for Multi-Faceted Efficient Adaptation of Large Models
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2312.00700