Skip to content

Latest commit

 

History

History
13 lines (4 loc) · 510 Bytes

README.md

File metadata and controls

13 lines (4 loc) · 510 Bytes

This is the work I have done on Model Compression during my time at Sony as an Intern.

The folder named CondenseNet contains training and testing related files for CondenseNet.

The folder named MobileNet contains training and testing related files for original MobileNets and MobileNetV2 combined with LGC(learned group convolution)/GC(group convolution).

The folder named ResultAnalysis contains matlab files for analyzing weights sparsity on expansion and projection layers of MobileNetV2_LGC/GC