Improved PyTorch implementation of https://github.com/mightydeveloper/Deep-Compression-PyTorch for'Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding' by Song Han, Huizi Mao, William J. Dally
Most of the code heavily borrowed from https://github.com/mightydeveloper/Deep-Compression-PyTorch
The configurations and running steps are referred to https://github.com/mightydeveloper/Deep-Compression-PyTorch
This code has been extended for the unstructured convolution transformation, which can conducted compression for linear and convolution layers. More construction such as Resnet and Darknet can be compressed by this code.