Comparing rewinding and fine-tuning in neural network pruning
Author(s)Renda, Alex(Alexander Dominic)
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science.
MetadataShow full item record
Many neural network pruning algorithms proceed in three steps: train the network to completion, remove unwanted structure to compress the network, and retrain the remaining structure to recover lost accuracy. The standard retraining technique, fine-tuning, trains the unpruned weights from their final trained values using a small fixed learning rate. In this thesis, I compare fine-tuning to alternative retraining techniques. Weight rewinding (as proposed by Frankle et al. (2019)), rewinds unpruned weights to their values from earlier in training and retrains them from there using the original training schedule. Learning rate rewinding (proposed in this thesis) trains the unpruned weights from their final values using the same learning rate schedule as weight rewinding. Both rewinding techniques outperform fine-tuning, forming the basis of a network-agnostic pruning algorithm that matches the accuracy and compression ratios of several more network-specific state-of-the-art techniques.
Thesis: S.M., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, May, 2020Cataloged from the official PDF of thesis.Includes bibliographical references (pages 40-44).
DepartmentMassachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Massachusetts Institute of Technology
Electrical Engineering and Computer Science.