diff --git a/stargan/attacks.py b/stargan/attacks.py
index 49a29ee..fa2acf0 100644
--- a/stargan/attacks.py
+++ b/stargan/attacks.py
@@ -7,7 +7,7 @@ import torch
 import torch.nn as nn
 
 class LinfPGDAttack(object):
-    def __init__(self, model=None, device=None, epsilon=0.05, k=20, a=0.01, feat = None):
+    def __init__(self, model=None, device=None, epsilon=0.05, k=1, a=0.05, feat = None):
         self.model = model
         self.epsilon = epsilon
         self.k = k
diff --git a/stargan/solver.py b/stargan/solver.py
index 7966142..76c97fe 100644
--- a/stargan/solver.py
+++ b/stargan/solver.py
@@ -569,7 +569,7 @@ class Solver(object):
                 save_image(self.denorm(x_concat.data.cpu()), result_path, nrow=1, padding=0)
                 print('Saved real and fake images into {}...'.format(result_path))
 
-    def test_attack_output(self):
+    def test_attack(self):
         """Translate images using StarGAN trained on a single dataset."""
 
         layer_dict = {0: 2, 1: 5, 2: 8, 3: 9, 4: 10, 5: 11, 6: 12, 7: 13, 8: 14, 9: 17, 10: 20, 11: None}
@@ -595,10 +595,10 @@ class Solver(object):
 
         for i, (x_real, c_org) in enumerate(data_loader):
             # Black image
-            # black = np.zeros((1,3,256,256))
-            # black = torch.FloatTensor(black).to(self.device)
+            black = np.zeros((1,3,256,256))
+            black = torch.FloatTensor(black).to(self.device)
 
-            black = torch.FloatTensor(torch.rand((1,3,256,256))).to(self.device)
+            # black = torch.FloatTensor(torch.rand((1,3,256,256))).to(self.device)
 
             # Prepare input images and target domain labels.
             x_real = x_real.to(self.device)
@@ -658,7 +658,7 @@ class Solver(object):
         print('{} images. L1 error: {}. L2 error: {}. L0 error: {}. L_-inf error: {}. Perceptual error: {}.'.format(n_samples, 
         l1_error / n_samples, l2_error / n_samples, l0_error / n_samples, min_dist / n_samples, perceptual_error / n_samples))
 
-    def test_attack(self):
+    def test_attack_feats(self):
         """Translate images using StarGAN trained on a single dataset."""
 
         layer_dict = {0: 2, 1: 5, 2: 8, 3: 9, 4: 10, 5: 11, 6: 12, 7: 13, 8: 14, 9: 17, 10: 20, 11: None}