/FL

Primary LanguagePython

Env setup

  • conda env create -f environment.yaml after changing the envs path to yours

Rest

Local

  • Server : cd rest/FL_Server && python3 manage.py runserver
  • Clients : cd rest && python3 run_local.py
    • args
      • ip : host ip (public ip for remote)
      • client : number of participating clients
      • exp : 1, 2, 3, 4
      • round : number of rounds
      • num : number of samples per client
      • delay : time delay until next command

Jetson

  • Server : cd rest/FL_Server && python3 manage.py runserver 0.0.0.0:20000
  • Clients :
    • docker pull crazyboy9103/jetson_fl:latest
    • docker run -ti -p 20000:20000 crazyboy9103/jetson_fl:latest
    • cd /FL/rest && python3 run_local.py --ip SERVER_IP --port 20000

TCP

Local

  • Server : cd socket && python3 multi_server.py

    • args
      • exp : 1, 2, 3, 4
      • num : number of samples per client
      • cli : number of participating clients
      • round : number of rounds
      • data : data mnist|cifar
      • host : host address (0.0.0.0 for remote)
      • port : port (forwarded port for remote)
      • epochs : number of local epochs
      • batch : size of batch

  • Clients : cd socket && python3 test_multi_clients.py

    • args
      • host : host address (public ip for remote)
      • port : port (forwarded port for remote)
      • cli : number of participating clients

Jetson

  • Server : python multi_server.py --host 0.0.0.0 --port 20000

  • Clients :

    • docker pull crazyboy9103/jetson_fl:latest
    • docker run -ti -p 20000:20000 crazyboy9103/jetson_fl:latest
    • cd /FL/socket && python3 test_multi_clients.py --host SERVER_IP --port 20000