Closed jaintj95 closed 4 years ago
Do you want to take this up?
Nope. My plate is already too full with the amount of work I have right now 😅 This is just a possible issue to tackle in the future if there's time.
Implementation looks complicated. Risk of breaking stuff if modified. No point in trying to fix something that works fine. Closing for now!
It seems like https://github.com/AlphaPav/DBA borrowed code from https://github.com/ebagdasa/backdoor_federated_learning
The later repo is almost 2 years old and I suspect that a lot of model configuration files maybe redundant now and we may be able to reduce complexity by directly using resnet models from torch.models