Eugene Bagdasarian comments

Results 15 comments of


                                            Eugene Bagdasarian

handling of compile_restricted optional params ``flags`` and ``dont_inherit``

is it related to this code: https://github.com/zopefoundation/RestrictedPython/blob/6c602c90cf8658cf1aaea660543228f127f6bbfd/src/RestrictedPython/compile.py#L70-L73 Why were these lines commented?

Question about parameter fl_eta in cifar_fed.yaml

yep, will check this week.

Questions about the low benign accuracy on CIFAR-10 and GTSRB dataset of Blind Backdoor

Hey Yiming, let me double check, I tried to simplify the API and might have broken something.

Questions about the low benign accuracy on CIFAR-10 and GTSRB dataset of Blind Backdoor

I think my config for cifar confused pre-trained models, but I think simply using a fixed scale and not MGDA will work better.

Questions about the low benign accuracy on CIFAR-10 and GTSRB dataset of Blind Backdoor

Hey, I actually never tested CIFAR as we did Imagenet in the paper so I didn't check it thoroughly, sorry. I believe it's fixed now. I added a threshold to...

Working with Hypriot 1.0 and kubernetes-on-arm 0.7.x

I didn't find brctl in .deb installation, should I install it manually? Same goes for wirte.sh

Backdoor accuracy can't reach 100%

Hey, I don't really support the code anymore, so it might not work well. Maybe try to test memorization first, i.e. mix test and train poison_images together.

dose I need to write the multi_mnist_params.yaml if I want to run the multi_mnist task?

Yeah, it's here: https://github.com/ebagdasa/backdoors101/blob/old_master/data/multi_mnist_loader.py but I didn't port it to the new redesigned version. Happy if you can make a pull request.

Questions about the low benign accuracy on CIFAR-10 and GTSRB dataset of Blind Backdoor

I guess you can just set it to 0.9

Questions about fl_task.update_global_model

I think it's updating the tensor of the model, let me know if it doesn't work