meta description: Making a deep convolutional neural network smaller and faster.
A user-friendly explanation how to compress CNN models - by removing full filters filters from a layer (GPU friendly, unlike sparse layers). L1-norm used for picking candidates for removal. Optimized MobileNet by 25%.
M. Braun, S. Krebs, F. Flohr, and D. Gavrila. (2018)cite arxiv:1805.07193Comment: Submitted to IEEE Trans. on Pattern Analysis and Machine Intelligence.
H. Zhao, O. Gallo, I. Frosio, and J. Kautz. (2015)cite arxiv:1511.08861Comment: This paper was published in IEEE Transactions on Computational Imaging on December 23, 2016.