It is a third-party implementation of the paper Drop an Octave: Reducing Spatial Redundancy in Convolutional Neural Networks with Octave Convolution
- Code for ResNet
- Code for the ablation study
- Trained models (rest exps)
- Thanks Mxnet implementation, our models are mostly based on their code