DyPrune: dynamic pruning rates for neural networks

Jonker, Richard A.A.; Poudel, Roshan; Fajarda, Olga; Oliveira, José Luís; Lopes, Rui Pedro; Matos, Sérgio

http://hdl.handle.net/10198/29602

Use this identifier to reference this record.

Name:	Description:	Size:	Format:
DyPrune Dynamic Pruning Rates for Neural Networks.pdf		414.73 KB	Adobe PDF	Download

Send Feedback

Authors

Abstract(s)

Neural networks have achieved remarkable success in various applications such as image classification, speech recognition, and natural language processing. However, the growing size of neural networks poses significant challenges in terms of memory usage, computational cost, and deployment on resource-constrained devices. Pruning is a popular technique to reduce the complexity of neural networks by removing unnecessary connections, neurons, or filters. In this paper, we present novel pruning algorithms that can reduce the number of parameters in neural networks by up to 98% without sacrificing accuracy. This is done by scaling the pruning rate of the models to the size of the model and scheduling the pruning to execute throughout the training of the model. Code related to this work is openly available.

Keywords

Machine learning Neural networks Pruning

URI

http://hdl.handle.net/10198/29602

Citation

Jonker, Richard A.A.; Poudel, Roshan; Fajarda, Olga; Oliveira, José Luís; Lopes, Rui Pedro; Matos, Sérgio (2023). DyPrune: dynamic pruning rates for neural networks. In Progress in Artificial Intelligence (EPIA). Cham: Springer. 14115. p. 146-157. ISBN 978-3-031-49007-1