Skip to content

Simple implementation of the FedAvg algorithm and experiments

Notifications You must be signed in to change notification settings

amir-thatoneguy/FedAvg_implementation

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 

Repository files navigation

FedAvg_implementation

Simple implementation of the FedAvg algorithm and experiments

How to Use

Change the experiment arguments (dictionaries) in the experiments section. It's easy to add other datasets or tasks to the code as well. The global section is to show that the FedAvg algorithm works as well as regular centralized training. I have not tested this code on larger models.

The main part is the training loops. In the local training loop function, a client trains their local model. In the federated training function, the overall training scheme is coordinated. Aggregation is done through the federated aggregation function, which is called at the end of every round in the federated training function.

About

Simple implementation of the FedAvg algorithm and experiments

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published