autopep8

dataloader/dataloader.py

@@ -68,8 +68,8 @@ def __getitem__(self, index):
             label = skimage.io.imread(self.target_images[index])
         if self.transform:
             image = self.transform(image)
-        image= torch.from_numpy(image)
-        label= torch.from_numpy(label)
+        image = torch.from_numpy(image)
+        label = torch.from_numpy(label)
         if torch.cuda.is_available():
             return {"Image": image.cuda(), "Label": label.cuda()}
         else:

hyperparams/hyperparams.py

@@ -1,6 +1,5 @@
 class hyperparameters:
     def __init__(self, **kwargs):
-        self.hyperparameters= dict()
+        self.hyperparameters = dict()
         for key, value in kwargs.items():
             self.hyperparameters.update({key: value})
-

loss/diceloss.py

@@ -1,14 +1,16 @@
 import torch
 import torch.nn as nn
 import torch.functional as F
+
+
 class Loss:
     @classmethod
     def dice_loss(self, pred, target, smooth=1.):
         pred = pred.contiguous()
         target = target.contiguous()
         intersection = (pred * target).sum(dim=2).sum(dim=2)
         loss = (1 - ((2. * intersection + smooth) /
-                    (pred.sum(dim=2).sum(dim=2) + target.sum(dim=2).sum(dim=2) + smooth)))
+                     (pred.sum(dim=2).sum(dim=2) + target.sum(dim=2).sum(dim=2) + smooth)))
         return loss.mean()
 
     def calc_loss(self, pred, target, metrics, bce_weight=0.5):

test/test_iter.py

@@ -15,5 +15,7 @@
     train_percentage=params.hyperparameters['train_percentage'])
 Train = TrainSet(Images.train_set, extension="tif", transform=None)
 Test = TestSet(Images.test_set, extension="tif", transform=None)
-for i in range(len(Train)):
-    print(Train[i].size())
+TrainLoder = DataLoader(Train, batch_size=4)
+for i, data in enumerate(TrainLoder, 0):
+    inputs, labels = data["Image"], data["Label"]
+    print(inputs.size())

train.py

@@ -9,12 +9,12 @@
 import torch
 from torchvision import transforms
 
-transforms_compose= transforms.Compose([transforms.ToTensor()])
+transforms_compose = transforms.Compose([transforms.ToTensor()])
 params = hyperparameters(train_percentage=0.6, batch_size=1, epoch=4)
 if torch.cuda.is_available():
-    net= UNeT(n_class=1).cuda()
+    net = UNeT(n_class=1).cuda()
 else:
-    net= net= UNeT(n_class=1)
+    net = net = UNeT(n_class=1)
 
 IMAGE_DIR = "/Users/madhav/DataSets/AerialImageDataset/train/images/*.tif"
 ANNOTATIONS_DIR = "/Users/madhav/DataSets/AerialImageDataset/train/gt/*.tif"
@@ -23,7 +23,10 @@
     Annotations=ANNOTATIONS_DIR,
     train_percentage=0.7)
 loss_val = Loss()
-Train = TrainSet(Images.train_set, extension="tif", transform=transforms_compose)
+Train = TrainSet(
+    Images.train_set,
+    extension="tif",
+    transform=transforms_compose)
 Test = TestSet(Images.test_set, extension="tif", transform=None)
 TrainLoder = DataLoader(
     Train,
@@ -39,7 +42,7 @@
     metrics = defaultdict()
     running_loss = 0.0
     for i, data in enumerate(TrainLoder, 0):
-        inputs, labels= data["Image"], data["Label"]
+        inputs, labels = data["Image"], data["Label"]
         #inputs, labels= inputs.permute(0, 3, 1, 2), labels.permute(0, 3, 1, 2)
         optimizer.zero_grad()
         outputs = net(inputs)
@@ -48,7 +51,7 @@
         optimizer.step()
         running_loss += loss.item()
         print('[%d, %5d] loss: %.3f' %
-        (epoch + 1, i + 1, running_loss / 2000))
+              (epoch + 1, i + 1, running_loss / 2000))
         running_loss = 0.
 
 print('Finished Training')