Closed hsndbk4 closed 4 years ago
Hi, the exact training hyperparameters can be found in Table 5, and attack parameters can be found in the caption of Table 7. This codebase mainly reproduces the results of our RobNets models. For baseline adversarial training, any standard codebase should be fine.
I have been trying to do adversarial training on MobileNetV2 for CIFAR10 using a standard codebase, and my natural/robust accuracies are no where near the reported numbers in Table 7 of the supplementary material. Could you please share the exact training hyperparameters for the MobileNetV2 model? or the model parameters if that is possible. Thanks.