Bi-directional Masks for Efficient N:M Sparse Training

TL;DR

This paper introduces Bi-Mask, a novel bi-directional masking technique that improves training efficiency for N:M sparse neural networks by enabling backward acceleration and maintaining high performance.

Contribution

The paper proposes a new bi-directional masking method with weight permutation to enhance N:M sparse training efficiency and performance.

Findings

Bi-Mask achieves better training acceleration than uni-directional masks.

Bi-Mask performs on par or better than methods without backward acceleration.

Experimental results validate the effectiveness of Bi-Mask in N:M sparse training.

Abstract

We focus on addressing the dense backward propagation issue for training efficiency of N:M fine-grained sparsity that preserves at most N out of M consecutive weights and achieves practical speedups supported by the N:M sparse tensor core. Therefore, we present a novel method of Bi-directional Masks (Bi-Mask) with its two central innovations in: 1) Separate sparse masks in the two directions of forward and backward propagation to obtain training acceleration. It disentangles the forward and backward weight sparsity and overcomes the very dense gradient computation. 2) An efficient weight row permutation method to maintain performance. It picks up the permutation candidate with the most eligible N:M weight blocks in the backward to minimize the gradient gap between traditional uni-directional masks and our bi-directional masks. Compared with existing uni-directional scenario that applies a transposable mask and enables backward acceleration, our Bi-Mask is experimentally demonstrated to be more superior in performance. Also, our Bi-Mask performs on par with or even better than methods that fail to achieve backward acceleration. Project of this paper is available at \url{https://github.com/zyxxmu/Bi-Mask}.