本文介绍了一种在移动云计算环境中，针对深度神经网络(DNN)的高效计算卸载框架。该框架考虑了移动设备的电池使用限制及云端服务器资源限制，通过整数线性规划(ILP)方法，实现DNN在移动设备与云端的高效协作。实验采用TensorFlow框架，NVIDIA Jetson TX2板作为移动设备，NVIDIA Tesla K40C为服务器平台。
Energy and Performance Efficient Computation Offloading for Deep Neural Networks in a Mobile Cloud Computing Environment
contribution
providing a framework to query DNNs efficiently on mobile devices collaborat- ing with cloud server in the presence of battery usage constraints and cloud server limited resources.
solution
ILP
实验设置
framework: tensorflow
mobile device: Jetson TX2 board developed by NVIDIA
server platform: NVIDIA®Tesla TMK40C
扫一扫
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
