arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2111.01697cs.LG
收藏

神经网络的低秩+稀疏张量压缩

Low-Rank+Sparse Tensor Compression for Neural Networks

University of California, Santa Barbara(加州大学圣巴巴拉分校);Facebook Inc.(Facebook公司)

Cole Hawkins, Haichuan Yang, Meng Li, Liangzhen Lai, Vikas Chandra

摘要:

低秩张量压缩已被提出作为一种有前途的方法,以减少神经网络部署在边缘设备上的内存和计算需求。张量压缩通过假设网络权重具有粗略的高阶结构来减少表示神经网络权重所需的参数的数量。这种粗结构假设已经被应用于压缩大型神经网络,如VGG和ResNet。然而,用于计算机视觉任务的现代最先进的神经网络(即MobileNet,EfficientNet)已经通过深度可分离卷积假设了粗糙的因子分解结构,使得纯张量分解成为一种不太有吸引力的方法。我们建议将低秩张量分解与稀疏修剪相结合,以便利用粗结构和细结构进行压缩。我们在SOTA架构(MobileNetv3,EfficientNet,视觉Transformer)中压缩权重,并将这种方法与稀疏修剪和张量分解进行比较。

英文摘要:

Low-rank tensor compression has been proposed as a promising approach to reduce the memory and compute requirements of neural networks for their deployment on edge devices. Tensor compression reduces the number of parameters required to represent a neural network weight by assuming network weights possess a coarse higher-order structure. This coarse structure assumption has been applied to compress large neural networks such as VGG and ResNet. However modern state-of-the-art neural networks for computer vision tasks (i.e. MobileNet, EfficientNet) already assume a coarse factorized structure through depthwise separable convolutions, making pure tensor decomposition a less attractive approach. We propose to combine low-rank tensor decomposition with sparse pruning in order to take advantage of both coarse and fine structure for compression. We compress weights in SOTA architectures (MobileNetv3, EfficientNet, Vision Transformer) and compare this approach to sparse pruning and tensor decomposition alone.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑