TinfoilHat0 / Defending-Against-Backdoors-with-Robust-Learning-Rate

The code of AAAI-21 paper titled "Defending against Backdoors in Federated Learning with Robust Learning Rate".

Home Page:https://ojs.aaai.org/index.php/AAAI/article/view/17118

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

about setup

b856741 opened this issue · comments

commented

hi,when I tried to run python federated.py --data=fmnist --local_ep=2 --bs=256 --num_agents=10 --rounds=200
there is a problem show up
err
do you have any ideas about it ?

Hi,

It looks like this error occurs due to tensorboard. Can you try commenting out tensorboard stuff, and try again?
If it works, I'd guess the error is due to some tensorboard version conflict between tensorflow and pytorch.

commented

Thanks a lot for replying.
I don't know if I commented the correct line
Try1
but after I commented this line, there's a new problem showed up.
try1err

You’ve to comment out all summarywriter events, not just the import.

commented

Hi,I tried to comment all summarywriter ,and writer,but it seems "writer" is a necessary part.
do you have else ideas? Thanks
try2
try2err

Just try setting writer=None, and comment out wherever writer is used maybe? The other solution is to simply solve the version conflict. You can try creating another environment without tensorflow, and just install pytorch and its version of tensorboard.

Closing as this is a package conflict