arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1703.03073cs.LGcs.CV
收藏

浮点权重定点激活的深度卷积神经网络推理

Deep Convolutional Neural Network Inference with Floating-point Weights and Fixed-point Activations

ARM Research(ARM研究院)

Liangzhen Lai, Naveen Suda, Vikas Chandra

摘要:

深度卷积神经网络(CNN)推理需要大量的内存和计算,这限制了其在嵌入式设备上的部署。为了在一定程度上缓解这些问题,先前的研究利用低精度定点数来表示CNN权重和激活。然而,固定点权重的最小所需数据精度在不同网络之间以及在同一网络的不同层之间变化。在这项工作中,我们建议使用浮点数表示的权重和定点数表示的激活。我们表明,对于相同的位宽,使用浮点表示的权重比定点表示更有效,并在流行的大规模CNN(如AlexNet,SqueezeNet,GoogLeNet和VGG-16)上进行了演示。我们还表明,这样的表示方案,使紧凑的硬件乘法和累加(MAC)单元设计。实验结果表明,该方案减少了高达36%的重量存储和功耗的硬件乘法器高达50%。

英文摘要:

Deep convolutional neural network (CNN) inference requires significant amount of memory and computation, which limits its deployment on embedded devices. To alleviate these problems to some extent, prior research utilize low precision fixed-point numbers to represent the CNN weights and activations. However, the minimum required data precision of fixed-point weights varies across different networks and also across different layers of the same network. In this work, we propose using floating-point numbers for representing the weights and fixed-point numbers for representing the activations. We show that using floating-point representation for weights is more efficient than fixed-point representation for the same bit-width and demonstrate it on popular large-scale CNNs such as AlexNet, SqueezeNet, GoogLeNet and VGG-16. We also show that such a representation scheme enables compact hardware multiply-and-accumulate (MAC) unit design. Experimental results show that the proposed scheme reduces the weight storage by up to 36% and power consumption of the hardware multiplier by up to 50%.

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。

↑