libowen424

can't set model.eval

3

When i run attack(), i need model.eval() and get loss to optime the adversarial example. However, it run error `RuntimeError: cudnn RNN backward can only be called in training mode`....

Thank you. May I ask if you have encountered a situation where lld_loss is consistently negative (such as an abnormal value of -190)? How should this be resolved?