Efficient Micro-Structured Weight Unification and Pruning for Neural Network Compression

TL;DR

This paper introduces a hardware-compatible micro-structured weight unification and pruning method for neural network compression, achieving high compression ratios and acceleration while maintaining task performance.

Contribution

It proposes a generalized micro-structured weight unification framework combined with ADMM-based training for efficient neural network compression.

Findings

Achieves state-of-the-art compression and acceleration results.

Maintains high task performance with low accuracy loss.

Compatible with various benchmark models and datasets.

Abstract

Compressing Deep Neural Network (DNN) models to alleviate the storage and computation requirements is essential for practical applications, especially for resource limited devices. Although capable of reducing a reasonable amount of model parameters, previous unstructured or structured weight pruning methods can hardly truly accelerate inference, either due to the poor hardware compatibility of the unstructured sparsity or due to the low sparse rate of the structurally pruned network. Aiming at reducing both storage and computation, as well as preserving the original task performance, we propose a generalized weight unification framework at a hardware compatible micro-structured level to achieve high amount of compression and acceleration. Weight coefficients of a selected micro-structured block are unified to reduce the storage and computation of the block without changing the neuron connections, which turns to a micro-structured pruning special case when all unified coefficients are set to zero, where neuron connections (hence storage and computation) are completely removed. In addition, we developed an effective training framework based on the alternating direction method of multipliers (ADMM), which converts our complex constrained optimization into separately solvable subproblems. Through iteratively optimizing the subproblems, the desired micro-structure can be ensured with high compression ratio and low performance degradation. We extensively evaluated our method using a variety of benchmark models and datasets for different applications. Experimental results demonstrate state-of-the-art performance.