Multi-Task Federated Learning for Personalised Deep Neural Networks in Edge Computing

阅读量：

1、torch中的函数名改变

将 set_deterministic 改为 use_deterministic_algorithms

2、运行命令

复制代码

    python main.py -dset mnist -alg fedavg -C 1.0 -B 150 -T 600 -E 2 -device cpu -W 200 -seed 2 -lr 0.1 -noisy_frac 0.1 -bn_private usyb
    
    
    AI写代码

复制代码

    python main.py -dset cifar10 -alg fedavg -C 1.0 - B 150 -T 600 -E 2 -device cpu -W 200 -seed 2 -lr 0.1 -noisy_frac 0.1 -bn_private usyb
    
    AI写代码

这些是重现表 2 和表 3 中结果的参数吗？如果是这样，我们使用 20 和 1 epoch 的批量大小，所以尝试：

复制代码

    python main.py -dset mnist -alg fedavg -C 1.0 -B 20 -T 500 -E 1 -device cpu -W 200 -seed 2 -lr 0.1 -noisy_frac 0.1 -bn_private usyb
    
    AI写代码

3、将库安装到指定环境

（conda无法使用国内镜像源时，使用pip安装到conda虚拟环境中。conda和pip语句安装的区别，conda list 和 pip list 显示结果不同。

目标路径为anaconda-envs-虚拟环境名-lib-python版本-site-packages）

复制代码

    pip install pickle --target=F:\Anaconda\envs\pytorch\Lib\site-packages -i http://mirrors.aliyun.com/pypi/simple scipy --trusted-host mirrors.aliyun.com
    
    AI写代码

4、progressbar版本问题需要增加初始化语句

复制代码

 # from progressbar import progressbar

    
 from progressbar import *
    
 progressbar = ProgressBar()
    
    
    
    
    AI写代码

全部评论 (0)

还没有任何评论哟~

Multi-Task Federated Learning for Personalised Deep Neural Networks in Edge Computing

1、torch中的函数名改变将setdeterministic改为usedeterministicalgorithms 2、运行命令 pythonmain.pydsetmnistalgfedavgC...

An Overview of Multi-Task Learning in Deep Neural Networks

目录摘要 1\.引言 2\.动机 3\.两种深度学习多任务学习方法 3.1.硬参数共享 3.2.软参数共享 4\.为什么多任务学习有效？ 4.1.隐含的数据增强 4.2.注意力聚焦 4.3.窃听 4...

Differentially Private Asynchronous Federated Learning for Mobile Edge Computing in Urban Informatic

DifferentiallyPrivateAsynchronousFederatedLearningforMobileEdgeComputinginUrbanInformatics阅读笔记文献背景及...

Decentralized Task Offloading in Edge Computing: A Multi-User Multi-Armed Bandit Approach

翻译：边缘计算中的分散式任务卸载：一种多用户多武装匪徒方法来源：IEEEINFOCOM2022IEEEConferenceonComputerCommunications 链接：Decentrali...

Multi-Task Deep Neural Networks for Natural Language Understanding阅读笔记

MTDNN Introduction 学习文本的向量空间表达对许多自然语言理解问题都很重要. 现在两个比较流行的方法是 multitasklearning languagemodelpretraini...

多任务学习-An Overview of Multi-Task Learning in Deep Neural Networks论文笔记

AnOverviewofMultiTaskLearninginDeepNeuralNetworks论文笔记概述：多任务学习有很多形式，如联合学习（JointLearning），自主学习（Learni...

Heterogeneous Multi-task Learning for Human Pose Estimation with Deep Convolutional Neural Network

HeterogeneousMultitaskLearningforHumanPoseEstimationwithDeepConvolutionalNeuralNetwork 论文题目Heterogen...

Learning Structured Sparsity in Deep Neural Networks

论文链接：<https://arxiv.org/pdf/1608.03665.pdf 代码链接：<https://github.com/wenwei202/caffe/tree/scnn 摘要计算资...

READ-2335 Toward Cleansing Backdoored Neural Networks in Federated Learning

论文名称TowardCleansingBackdooredNeuralNetworksinFederatedLearning 作者ChenWu,XianYang,SencunZhu,Prasenjit...

Cross-stitch Networks for Multi-task Learning译文_

目录 1.简介 1.1多任务共享：一项实证研究 2.相关工作 3十字绣网络 3.1.分割架构 3.2.统一拆分架构 3.3.十字绣单元 4.十字拼接的设计决策 5.消融分析 5.1.初始化十字绣单元的...

是否确定退出登录?

Multi-Task Federated Learning for Personalised Deep Neural Networks in Edge Computing

1、torch中的函数名改变

2、运行命令

3、将库安装到指定环境

4、progressbar版本问题需要增加初始化语句

全部评论 (0)

相关文章推荐

Multi-Task Federated Learning for Personalised Deep Neural Networks in Edge Computing

An Overview of Multi-Task Learning in Deep Neural Networks

Differentially Private Asynchronous Federated Learning for Mobile Edge Computing in Urban Informatic

Decentralized Task Offloading in Edge Computing: A Multi-User Multi-Armed Bandit Approach

Multi-Task Deep Neural Networks for Natural Language Understanding阅读笔记

多任务学习-An Overview of Multi-Task Learning in Deep Neural Networks论文笔记

Heterogeneous Multi-task Learning for Human Pose Estimation with Deep Convolutional Neural Network

Learning Structured Sparsity in Deep Neural Networks

READ-2335 Toward Cleansing Backdoored Neural Networks in Federated Learning

Cross-stitch Networks for Multi-task Learning译文_