arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2103.12896cs.CVcs.LGeess.IV
收藏

SETGAN:移动平台上图像应用的规模和能量权衡GANs

SETGAN: Scale and Energy Trade-off GANs for Image Applications on Mobile Platforms

School of EECS, Washington State University(华盛顿州立大学电气与计算机工程学院)

Nitthilan Kannappan Jayakodi, Janardhan Rao Doppa, Partha Pratim Pande

摘要:

我们考虑使用生成对抗网络(GANs)在移动平台上实现照片级真实感无条件图像生成(生成高质量,多样化的样本,携带与图像相同的视觉内容)的任务。在本文中,我们提出了一种新的方法来权衡图像生成精度的GAN的能量消耗(计算)在运行时称为规模能量权衡GAN(SET GAN)。GANs通常需要很长时间来进行训练,并消耗大量内存,因此很难在边缘设备上运行。图像生成任务的SETGAN背后的关键思想是,对于给定的输入图像,我们在远程服务器上训练一个GAN,并在边缘设备上使用经过训练的模型。我们使用SinGAN,这是一种单图像无条件生成模型,它包含一个完全卷积的GANs金字塔,每个GAN负责学习图像不同尺度下的补丁分布。在训练过程中,我们确定给定输入图像的最佳尺度数和目标边缘设备的能量约束。结果表明,利用SETGAN独特的基于客户端-服务器的架构,我们能够在损失3%至12% SSIM准确性的情况下实现56%的能量增益。此外,通过并行多尺度训练,我们在服务器上获得了大约4倍的训练时间增益。

英文摘要:

We consider the task of photo-realistic unconditional image generation (generate high quality, diverse samples that carry the same visual content as the image) on mobile platforms using Generative Adversarial Networks (GANs). In this paper, we propose a novel approach to trade-off image generation accuracy of a GAN for the energy consumed (compute) at run-time called Scale-Energy Tradeoff GAN (SETGAN). GANs usually take a long time to train and consume a huge memory hence making it difficult to run on edge devices. The key idea behind SETGAN for an image generation task is for a given input image, we train a GAN on a remote server and use the trained model on edge devices. We use SinGAN, a single image unconditional generative model, that contains a pyramid of fully convolutional GANs, each responsible for learning the patch distribution at a different scale of the image. During the training process, we determine the optimal number of scales for a given input image and the energy constraint from the target edge device. Results show that with SETGAN's unique client-server-based architecture, we were able to achieve a 56% gain in energy for a loss of 3% to 12% SSIM accuracy. Also, with the parallel multi-scale training, we obtain around 4x gain in training time on the server.

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。

↑