search

cat-6sephiruth / 21_CSRC

0 stars 0 forks source link

Same problem of "backdoor federated learning" #1

Open Growdy opened 3 years ago

Growdy commented 3 years ago

Sorry for putting the issue here, but I can't contact you by email. I saw your issue under the code of "how to backdoor federated learning".　I meet the same problem as you, but I can't solve it. I don't know how to tweak the params file. Can you help me?

current path: saved_models/model_image_May.03_22.55.09 Loading data Files already downloaded and verified We use following environment for graphs: ppdl_experiment Selected models: [65, 7, 66, 27, 80, 93, 50, 34, 4, 52] There are 0 adversaries in the training. time spent on training: 36.929800510406494 Traceback (most recent call last): File "training.py", line 671, in weight_accumulator=weight_accumulator, epoch=epoch) File "D:\should\backdoor_federated_learning-master\backdoor_federated_learning-master\helper.py", line 211, in average_shrinkmodels data.add(update_per_layer) RuntimeError: result type Float can't be cast to the desired output type Long

cat-6sephiruth commented 3 years ago

Hello, the progress has been going on for a long time, so I think you need to tell me in detail what the symptoms are. call the 6sephiruth@gmail.com In addition, it would be better if you refer to this site. [(https://github.com/AI-secure/DBA)] Upgrade the paper above? One paper, and it works fine with that code. I didn't understand what your problem was. If you tell me the details, I'll help you.