Model training on Python (using Keras/TF)
Network simulation using ns3
Device constraints like available RAM, GPU applied
Simulation1 : CIFAR dataset across 10 clients Simulation 2 : Heterogenous data split and training across 200 clients
- Run the
run setup.sh
file to setup everything. - If gpu is not required then set the gpu option in
config.yml
as False. - If gpu is required then first set the gpu option in
config.yml
as True and then also run thegpu_setup.sh
file.
Check that all the above installation steps are completed before running main.py.
Everything can be executed form the main.py
. Requires passing ns3 home path through command line input. All configurations maintained in config.yml
.
Code associated with paper [https://arxiv.org/abs/2405.17839]