EigenDamage: Structured Pruning in the Kronecker-Factored Eigenbasis

Abstract

Reducing the test time resource requirements of a neural network while preserving test accuracy is crucial for running inference on resource-constrained devices. To achieve this goal, we introduce a novel network reparameterization based on the Kronecker-factored eigenbasis (KFE), and then apply Hessian-based structured pruning methods in this basis. As opposed to existing Hessian-based pruning algorithms which do pruning in parameter coordinates, our method works in the KFE where different weights are approximately independent, enabling accurate pruning and fast computation. We demonstrate empirically the effectiveness of the proposed method through extensive experiments. In particular, we highlight that the improvements are especially significant for more challenging datasets and networks. With negligible loss of accuracy, an iterative-pruning version gives a 10$\times$ reduction in model size and a 8$\times$ reduction in FLOPs on wide ResNet32.

ICML 2019

Cited by in corpus (6)

• End-to-End Supermask Pruning: Learning to Prune Image Captioning Models
• Single Shot Structured Pruning Before Training
• Convolutional Neural Network Pruning with Structural Redundancy Reduction
• Traces of Class/Cross-Class Structure Pervade Deep Learning Spectra
• Neural Pruning via Growing Regularization
• Deep Neural Compression Via Concurrent Pruning and Self-Distillation