Jack-lx-jiang

To train JARN-AT1, I tried to perturb the training date by FGSM in train_jarn.py: ``` attack = LinfPGDAttack(model, 8.0, 1, 8.0, False, 'xent', 'cifar10') ``` ``` if img_random_pert: x_batch =...

0 loss when using ssd_pascal with my own dataset

10

Hi, When I used ssd to train my own dataset, I always got a 0 loss at the beginning and then the loss bounce around zero every iteration(sometimes even got...