Enregistré dans:
Détails bibliographiques
Auteurs principaux: Menghani, Gaurav, Kumar, Ravi, Kumar, Sanjiv
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:https://arxiv.org/abs/2411.07501
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866916808168046592
author Menghani, Gaurav
Kumar, Ravi
Kumar, Sanjiv
author_facet Menghani, Gaurav
Kumar, Ravi
Kumar, Sanjiv
contents One of the core pillars of efficient deep learning methods is architectural improvements such as the residual/skip connection, which has led to significantly better model convergence and quality. Since then the residual connection has become ubiquitous in not just convolutional neural networks but also transformer-based architectures, the backbone of LLMs. In this paper we introduce Learned Augmented Residual Layer (LAuReL) -- a novel generalization of the canonical residual connection -- with the goal to be an in-situ replacement of the latter while outperforming on both model quality and footprint metrics. Our experiments show that using LAuReL can help boost performance for both vision and language models. For example, on the ResNet-50, ImageNet 1K task, it achieves 60% of the gains from adding an extra layer, while only adding 0.003% more parameters, and matches it while adding 2.6 times fewer parameters. Similarly, when pre-training 1B and 4B parameter LLMs, LAuReL improves performance on a variety of challenging downstream evaluation tasks by 2.54% to 20.05%, while adding only 0.012% and 0.1% additional parameters, respectively.
format Preprint
id arxiv_https___arxiv_org_abs_2411_07501
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle LAuReL: Learned Augmented Residual Layer
Menghani, Gaurav
Kumar, Ravi
Kumar, Sanjiv
Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
One of the core pillars of efficient deep learning methods is architectural improvements such as the residual/skip connection, which has led to significantly better model convergence and quality. Since then the residual connection has become ubiquitous in not just convolutional neural networks but also transformer-based architectures, the backbone of LLMs. In this paper we introduce Learned Augmented Residual Layer (LAuReL) -- a novel generalization of the canonical residual connection -- with the goal to be an in-situ replacement of the latter while outperforming on both model quality and footprint metrics. Our experiments show that using LAuReL can help boost performance for both vision and language models. For example, on the ResNet-50, ImageNet 1K task, it achieves 60% of the gains from adding an extra layer, while only adding 0.003% more parameters, and matches it while adding 2.6 times fewer parameters. Similarly, when pre-training 1B and 4B parameter LLMs, LAuReL improves performance on a variety of challenging downstream evaluation tasks by 2.54% to 20.05%, while adding only 0.012% and 0.1% additional parameters, respectively.
title LAuReL: Learned Augmented Residual Layer
topic Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2411.07501