Skip to content

Files

Latest commit

 

History

History
5 lines (3 loc) · 514 Bytes

interpret-cnn-compress.md

File metadata and controls

5 lines (3 loc) · 514 Bytes

1. Structural compression of convolutional neural networks based on greedy filter pruning - [arXiv]

2. Interpreting Convolutional Neural Networks Through Compression - [arXiv]

The author proposed classification accuracy reduction (CAR). In the CAR structural compression, the filter with the least effect on the classification accuracy gets pruned in each iteration. Afterwards, a fine tuning process is needed to get better accuracy.