Skip to content

scratches for now to validate feasibility of a 100 perc. decentralized federated learning method in the presence of byzantine participants, using unbalanced data sets

Notifications You must be signed in to change notification settings

andrasfe/decfed

Repository files navigation

WIP A completely decentralized Federated Learning approach

scratches for now to validate feasibility of a completely decentralized federated learning method in the presence of byzantine participants, using unbalanced data sets

currently implementing logic from this paper. Forked a Kaggle kernel.

Initially all simple threaded, no use of tensors, until I can validate functionality.

About

scratches for now to validate feasibility of a 100 perc. decentralized federated learning method in the presence of byzantine participants, using unbalanced data sets

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published