Model Preserving Compression for Neural Networks

Jerry Chee, Megan Renz, Anil Damle +1

After training complex deep learning models, a common task is to compress the model to reduce compute and storage demands. When compressing, it is desirable to preserve the origina…