Ashish Patel 🇮🇳’s Post

Day-16 Computer Vision Learning SOTA RepVGG: Making VGG-style ConvNets Great Again Published on 11 January 2021 --------------------------------------------------------------- 𝗔𝗺𝗮𝘇𝗶𝗻𝗴 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 : https://lnkd.in/eaU_EuD 𝗣𝘆𝘁𝗼𝗿𝗰𝗵: https://lnkd.in/eNkFnB7 𝗠𝗲𝗴𝘃𝗶𝗶 𝗠𝗼𝗱𝗲𝗹 𝗥𝗲𝗹𝗲𝗮𝘀𝗲 𝘀𝗼𝗼𝗻 : https://lnkd.in/e_Ju3pT ------------------------------------------------------------------- 𝗜𝗠𝗣𝗢𝗥𝗧𝗔𝗡𝗖𝗘 ♦ This architecture of the 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 favourable accuracy-speed trade-off compared to the state-of-the-art models like EfficientNet and RegNet. Note: More in Comments #innovation #artificialintelligence #computervision

  • diagram, schematic

Re-param for Plain Inference-time Model

  • No alternative text description for this image

Performance Comparision with Other Model

  • No alternative text description for this image

𝗥𝗲𝗽𝗩𝗚𝗚 𝗵𝗮𝘀 𝘁𝗵𝗲 𝗳𝗼𝗹𝗹𝗼𝘄𝗶𝗻𝗴 𝗮𝗱𝘃𝗮𝗻𝘁𝗮𝗴𝗲𝘀. • The model has a VGG-like plain (a.k.a. feed-forward) topology 1 without any branches. I.e., every layer takes the output of its only preceding layer as input and feeds the output into its only following layer. • The model’s body uses only 3 × 3 conv and ReLU. • The concrete architecture (including the specific depth and layer widths) is instantiated with no automatic search, manual refinement, compound scaling, nor other heavy designs.

  • No alternative text description for this image
See more comments

To view or add a comment, sign in

Explore content categories