FedPAW: Federated Learning with Personalized Aggregation Weights for Urban Vehicle Speed Prediction

This project PFLlibVSP is based on the open source project PFLlib development.

Environments

Install CUDA.

Install conda and activate conda.

conda env create -f env_cuda_latest.yaml # You may need to downgrade the torch using pip to match CUDA version

How to start simulating (examples for FedAvg)

• Create proper environments (see Environments).

• Download this project to an appropriate location using git.

  git clone https://github.com/heyuepeng/PFLlibVSP.git

• Run evaluation:

  cd ./system
  python main.py -data driving -algo FedAvg -gr 300 -did 0 # using the driving dataset, the FedAvg algorithm, train 300 rounds

  Or you can uncomment the lines you need in ./system/examples.sh and run:

  cd ./system
  sh examples.sh

Note: The hyper-parameters have not been tuned for the algorithms. The values in ./system/examples.sh are just examples. You need to tune the hyper-parameters by yourself.

Datasets

Our dataset CarlaVSP is available from that link: https://pan.baidu.com/s/1qs8fxUvSPERV3C9i6pfUIw?pwd=tl3e.