Federated-Learning-PyTorch
Federated-Learning-PyTorch copied to clipboard
Why the test accuracy of CIFAR is so low ?
Even for the base line , the accuray is around 30%
Same issue. Is there any update?
I think for me, it was CIFAR dataset not MNIST. It was my mistake. (updated)
I hava the same question on this, I change some attributes,it archived 60%.Maybe use DNN will be better(hadn't test)
I hava the same question on this, I change some attributes,it archived 60%.Maybe use DNN will be better(hadn't test)
60% seems to be much better. Could you share which attributes you've changed? Thanks