Model Compression

We apply model compression techniques such as pruning and quantization to deep learning image compression models and find that pruning affects different components of these models differently.

This project focuses on memory compression techniques for image compression algorithms due to the high cost of transferring data and computation between different units.