Pytorch implementation of LRMF:Filter Pruning via Learned Representation Median in the Frequency Domain.
By running original_train.py, you can get the results of baseline models without pruning. The pruning_cifar10.py is used to implement our pruning method. The models folder contains all the necessary code for the network definition. The functions folder contains supporting code to reproduce our work.
This code is implemented based on CVPR 2019 Oral 'Filter Pruning via Geometric Median for Deep Convolutional Neural Networks Acceleration'