Merge e8e48a2070 into 4180346273

src/data/dataset.py

@@ -30,10 +30,10 @@ class InpaintingData(Dataset):
             transforms.ColorJitter(0.05, 0.05, 0.05, 0.05),
             transforms.ToTensor()])
         self.mask_trans = transforms.Compose([
-            transforms.Resize(args.image_size, interpolation=transforms.InterpolationMode.NEAREST),
+            transforms.Resize(args.image_size),
             transforms.RandomHorizontalFlip(),
             transforms.RandomRotation(
-                (0, 45), interpolation=transforms.InterpolationMode.NEAREST),
+                (0, 45)),
         ])
 
         
@@ -77,4 +77,4 @@ if __name__ == '__main__':
     data = InpaintingData(args)
     print(len(data), len(data.mask_path))
     img, mask, filename = data[0]
-    print(img.size(), mask.size(), filename)
+    print(img.size(), mask.size(), filename)

src/test.py

@@ -33,9 +33,9 @@ def main_worker(args, use_gpu=True):
     # prepare dataset
     image_paths = []
     for ext in ['.jpg', '.png']: 
-        image_paths.extend(glob(os.path.join(args.dir_image, '*'+ext)))
+        image_paths.extend(glob(os.path.join(args.dir_test, '*'+ext)))
     image_paths.sort()
-    mask_paths = sorted(glob(os.path.join(args.dir_mask, '*.png')))
+    mask_paths = sorted(glob(os.path.join(args.dir_mask,args.mask_type,'*.png')))
     os.makedirs(args.outputs, exist_ok=True)
     
     # iteration through datasets

src/utils/option.py

@@ -9,8 +9,8 @@ parser.add_argument('--dir_mask', type=str, default='../../dataset',
                     help='mask dataset directory')
 parser.add_argument('--data_train', type=str, default='places2',
                     help='dataname used for training')
-parser.add_argument('--data_test', type=str, default='places2',
+parser.add_argument('--dir_test', type=str, default='../datasets/test_imgs/',
-                    help='dataname used for testing')
+                    help='test image dataset directory')
 parser.add_argument('--image_size', type=int, default=512,
                     help='image size used during training')
 parser.add_argument('--mask_type', type=str, default='pconv',
@@ -93,4 +93,4 @@ losses = list(args.rec_loss.split('+'))
 args.rec_loss = {}
 for l in losses: 
     weight, name = l.split('*')
-    args.rec_loss[name] = float(weight)
+    args.rec_loss[name] = float(weight)