SOTAVerified

RepVGG: Making VGG-style ConvNets Great Again

2021-01-11CVPR 2021Code Available2· sign in to hype

Xiaohan Ding, Xiangyu Zhang, Ningning Ma, Jungong Han, Guiguang Ding, Jian Sun

arXivPDF

Code Available — Be the first to reproduce this paper.

Reproduce

Code

Abstract

We present a simple but powerful architecture of convolutional neural network, which has a VGG-like inference-time body composed of nothing but a stack of 3x3 convolution and ReLU, while the training-time model has a multi-branch topology. Such decoupling of the training-time and inference-time architecture is realized by a structural re-parameterization technique so that the model is named RepVGG. On ImageNet, RepVGG reaches over 80% top-1 accuracy, which is the first time for a plain model, to the best of our knowledge. On NVIDIA 1080Ti GPU, RepVGG models run 83% faster than ResNet-50 or 101% faster than ResNet-101 with higher accuracy and show favorable accuracy-speed trade-off compared to the state-of-the-art models like EfficientNet and RegNet. The code and trained models are available at https://github.com/megvii-model/RepVGG.

Tasks

GPUImage ClassificationSemantic Segmentation

Benchmark Results

DatasetModelMetricClaimedVerifiedStatus
ImageNetRepVGG-B2Top 1 Accuracy78.78—Unverified
ImageNetRepVGG-B2g4Top 1 Accuracy78.5—Unverified

Reproductions