narugo1992
dev(narugo): init application
65abdbf
import torch
from torch.cuda import amp
from .base import Attacker
class FGSM(Attacker):
def __init__(self, model, img_transform=(lambda x: x, lambda x: x), use_amp=False):
super().__init__(model, img_transform)
self.use_amp = use_amp
if use_amp:
self.scaler = amp.GradScaler()
def set_para(self, eps=8, alpha=lambda: 8, **kwargs):
super().set_para(eps=eps, alpha=alpha, **kwargs)
def step(self, images, labels, loss):
with amp.autocast(enabled=self.use_amp):
images.requires_grad = True
outputs = self.model(images).logits
self.model.zero_grad()
cost = loss(outputs, labels)
if self.use_amp:
self.scaler.scale(cost).backward()
else:
cost.backward()
adv_images = (images + self.alpha() * images.grad.sign()).detach_()
eta = torch.clamp(adv_images - self.ori_images, min=-self.eps, max=self.eps)
images = self.img_transform[0](
torch.clamp(self.img_transform[1](self.ori_images + eta), min=0, max=255).detach_())
return images
def attack(self, images, labels):
# images = deepcopy(images)
# self.ori_images = deepcopy(images)
self.model.eval()
images = self.forward(self, images, labels)
self.model.zero_grad()
self.model.train()
return images