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.
翻译:He等人设计的有影响力的残余网络仍然是许多科学出版物中的金标准架构,通常是作为研究中的默认架构,或者作为提出新架构时的基线;然而,自2015年ResNet架构启动以来,在培训神经网络的最佳做法方面取得了显著进展,新优化和数据增强提高了培训食谱的有效性。在本文件中,我们重新评估了Vanilla ResNet-50的绩效,在培训时采用了一种整合这些进步的程序。我们在Timm 开源图书馆中分享了竞争性培训设置和预先培训的模式,希望它们能成为未来工作的更好的基线。例如,随着我们要求更高的培训环境,Vanilla ResNet-50在没有额外数据或蒸馏的情况下,在关于图像网络价值的第224x224号决议中达到了80.4%的最高精度。我们还报告了与我们培训程序通用模式的绩效。