diff --git a/examples/auto_encoder/train.py b/examples/auto_encoder/train.py
index 1c3ab3b..5acb557 100644
--- a/examples/auto_encoder/train.py
+++ b/examples/auto_encoder/train.py
@@ -37,6 +37,9 @@ def get_argparser():
 
 
 def main():
+    if not os.path.exists('results'):
+        os.mkdir('results')
+
     opts = get_argparser().parse_args()
 
     # dataset
@@ -58,7 +61,7 @@ def main():
 
     val_loader = DataLoader(
         ImageDataset(root='datasets/CLIC/valid', transform=val_transform),
-        batch_size=opts.batch_size, shuffle=False, num_workers=0)
+        batch_size=1, shuffle=False, num_workers=0)
 
     print("Train set: %d, Val set: %d" %
           (len(train_loader.dataset), len(val_loader.dataset)))
@@ -105,7 +108,7 @@ def main():
         # =====  Validation  =====
         print("Val...")
         best_score = 0.0
-        cur_score = test(opts, model, val_loader)
+        cur_score = test(opts, model, val_loader, cur_epoch)
         print("%s = %.6f" % (opts.loss_type, cur_score))
         # =====  Save Best Model  =====
         if cur_score > best_score:  # save best model
@@ -114,7 +117,10 @@ def main():
             print("Best model saved as best_model.pt")
 
 
-def test(opts, model, val_loader):
+def test(opts, model, val_loader, epoch):
+    save_dir = os.path.join('results', 'epoch_%d' % epoch)
+    if not os.path.exists(save_dir):
+        os.mkdir(save_dir)
     model.eval()
     cur_score = 0.0
 
@@ -124,11 +130,10 @@ def test(opts, model, val_loader):
         for i, (images, ) in enumerate(val_loader):
             outputs = model(images)
             # save the first reconstructed image
-            if i == 20:
-                Image.fromarray((outputs*255).squeeze(0).detach().numpy().astype(
-                    'uint8').transpose(1, 2, 0)).save('recons_%s.png' % (opts.loss_type))
             cur_score += metric(outputs, images, data_range=1.0)
+            Image.fromarray((outputs*255).squeeze(0).detach().numpy().astype('uint8').transpose(1, 2, 0)).save(os.path.join(save_dir, 'recons_%s_%d.png' % (opts.loss_type, i)))
         cur_score /= len(val_loader.dataset)
+
     return cur_score