Description
RepVGG is a VGG-style convolutional architecture. It 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.
Papers Using This Method
On the Vulnerability of Skip Connections to Model Inversion Attacks2024-09-03Oracle Bone Script Similiar Character Screening Approach Based on Simsiam Contrastive Learning and Supervised Learning2024-08-13RepVGG-GELAN: Enhanced GELAN with VGG-STYLE ConvNets for Brain Tumour Detection2024-05-06Outlier-Aware Training for Low-Bit Quantization of Structural Re-Parameterized Networks2024-02-11NeRCC: Nested-Regression Coded Computing for Resilient Distributed Prediction Serving Systems2024-02-06Detection of Small Targets in Sea Clutter Based on RepVGG and Continuous Wavelet Transform2023-11-14UNISOUND System for VoxCeleb Speaker Recognition Challenge 20232023-08-24A region and category confidence-based multi-task network for carotid ultrasound image segmentation and classification2023-07-02Make RepVGG Greater Again: A Quantization-aware Approach2022-12-03Artificial Intelligence for Automatic Detection and Classification Disease on the X-Ray Images2022-11-14RMNet: Equivalently Removing Residual Connection from Networks2021-11-01Fast query-by-example speech search using separable model2021-09-18RepVGG: Making VGG-style ConvNets Great Again2021-01-11