ResNet strikes back: An improved training procedure in timm

Abstract

The influential Residual Networks designed by He et al. remain the gold-standard architecture in numerous scientific publications. They typically serve as the default architecture in studies, or as baselines when new architectures are proposed. Yet there has been significant progress on best practices for training neural networks since the inception of the ResNet architecture in 2015. Novel optimization & data-augmentation have increased the effectiveness of the training recipes. In this paper, we re-evaluate the performance of the vanilla ResNet-50 when trained with a procedure that integrates such advances. We share competitive training settings and pre-trained models in the timm open-source library, with the hope that they will serve as better baselines for future work. For instance, with our more demanding training setting, a vanilla ResNet-50 reaches 80.4% top-1 accuracy at resolution 224x224 on ImageNet-val without extra data or distillation. We also report the performance achieved with popular models with our training procedure.

Full paper on arXiv

Details

arXiv identifier

2110.00476

Published

2021-10-01

Authors

Ross Wightman, Hugo Touvron, Hervé Jégou

Models That Cite This Paper

• Described bymobilenetv3_small_100.lamb_in1k
• Described byefficientnet_b3.ra2_in1k
• Described byresnet18.a1_in1k
• Described byresnet50.a1_in1k
• Described byresnet18.a3_in1k
• Described byefficientnet_b0.ra_in1k
• Described byresnet34.a1_in1k
• Described bywide_resnet50_2.racm_in1k
• Described byresnet50_gn.a1h_in1k
• Described bydensenet121.ra_in1k
• Described byresnet50d.ra2_in1k