working code committed

Author: sohaib023

eval.py

@@ -1,92 +1,99 @@
-import os
-import argparse
-
-import cv2
-import numpy as np
-
-import torch
-from torch.utils.data import DataLoader
-
-from siamese import SiameseNetwork
-from libs.dataset import Dataset
-
-if __name__ == "__main__":
-    parser = argparse.ArgumentParser()
-
-    parser.add_argument(
-        '--val_path',
-        type=str,
-        help="Path to directory containing validation dataset.",
-        default="../dataset/test"
-    )
-    parser.add_argument(
-        '-o',
-        '--out_path',
-        type=str,
-        help="Path for outputting model weights and tensorboard summary.",
-        default="output/images"
-    )
-    parser.add_argument(
-        '-c',
-        '--checkpoint',
-        type=str,
-        help="Path to model to be used for inference.",
-        default="output/epoch_200.pth"
-    )
-
-    args = parser.parse_args()
-
-    os.makedirs(args.out_path, exist_ok=True)
-
-    device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
-    
-    val_dataset     = Dataset(args.val_path, shuffle_pairs=False, augment=False, testing=True)
-    val_dataloader   = DataLoader(val_dataset, batch_size=1)
-
-    model = SiameseNetwork()
-    model.to(device)
-
-    checkpoint = torch.load(args.checkpoint)
-    model.load_state_dict(checkpoint['model_state_dict'])
-
-    model.eval()
-
-    losses = []
-    correct = 0
-    total = 0
-
-    inv_transform = transforms.Compose([ transforms.Normalize(mean = [ 0., 0., 0. ],
-                                                         std = [ 1/0.229, 1/0.224, 1/0.225 ]),
-                                    transforms.Normalize(mean = [ -0.485, -0.456, -0.406 ],
-                                                         std = [ 1., 1., 1. ]),
-                                   ])
-    
-    for i, ((img1, img2), y, (class1, class2)) in enumerate(val_dataloader):
-        print("[{} / {}]".format(i, len(val_dataloader)))
-
-        img1, img2, y = map(lambda x: x.to(device), [img1, img2, y])
-
-        prob = model(img1, img2)
-        loss = criterion(prob, y)
-
-        losses.append(loss.item())
-        correct += torch.count_nonzero(y == (prob > 0.5)).item()
-        total += len(y)
-
-        fig = plt.figure("class1={}\tclass2={}".format(class1, class2), figsize=(4, 2))
-        plt.suptitle("cls1={}  conf={:.2f}  cls2={}".format(class1, prob[0], class2))
-
-        # show first image
-        ax = fig.add_subplot(1, 2, 1)
-        plt.imshow(inv_transform(img1[0]), cmap=plt.cm.gray)
-        plt.axis("off")
-
-        # show the second image
-        ax = fig.add_subplot(1, 2, 2)
-        plt.imshow(inv_transform(img2[0]), cmap=plt.cm.gray)
-        plt.axis("off")
-
-        # show the plot
-        plt.savefig(os.path.join(args.checkpoint, 'images/{}.png').format(i))
-
+import os
+import argparse
+
+import cv2
+import numpy as np
+import matplotlib.pyplot as plt
+
+import torch
+from torch.utils.data import DataLoader
+from torchvision import transforms
+
+from siamese import SiameseNetwork
+from libs.dataset import Dataset
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+
+    parser.add_argument(
+        '--val_path',
+        type=str,
+        help="Path to directory containing validation dataset.",
+        default="../dataset/test"
+    )
+    parser.add_argument(
+        '-o',
+        '--out_path',
+        type=str,
+        help="Path for outputting model weights and tensorboard summary.",
+        default="output/images"
+    )
+    parser.add_argument(
+        '-c',
+        '--checkpoint',
+        type=str,
+        help="Path to model to be used for inference.",
+        default="output/epoch_200.pth"
+    )
+
+    args = parser.parse_args()
+
+    os.makedirs(args.out_path, exist_ok=True)
+
+    device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+    
+    val_dataset     = Dataset(args.val_path, shuffle_pairs=False, augment=False, testing=True)
+    val_dataloader   = DataLoader(val_dataset, batch_size=1)
+
+    model = SiameseNetwork()
+    model.to(device)
+    criterion = torch.nn.BCELoss()
+
+    checkpoint = torch.load(args.checkpoint)
+    model.load_state_dict(checkpoint['model_state_dict'])
+
+    model.eval()
+
+    losses = []
+    correct = 0
+    total = 0
+
+    inv_transform = transforms.Compose([ transforms.Normalize(mean = [ 0., 0., 0. ],
+                                                         std = [ 1/0.229, 1/0.224, 1/0.225 ]),
+                                    transforms.Normalize(mean = [ -0.485, -0.456, -0.406 ],
+                                                         std = [ 1., 1., 1. ]),
+                                   ])
+    
+    for i, ((img1, img2), y, (class1, class2)) in enumerate(val_dataloader):
+        print("[{} / {}]".format(i, len(val_dataloader)))
+
+        img1, img2, y = map(lambda x: x.to(device), [img1, img2, y])
+        class1 = class1[0]
+        class2 = class2[0]
+
+        prob = model(img1, img2)
+        loss = criterion(prob, y)
+
+        losses.append(loss.item())
+        correct += torch.count_nonzero(y == (prob > 0.5)).item()
+        total += len(y)
+
+        fig = plt.figure("class1={}\tclass2={}".format(class1, class2), figsize=(4, 2))
+        plt.suptitle("cls1={}  conf={:.2f}  cls2={}".format(class1, prob[0][0].item(), class2))
+
+        img1 = inv_transform(img1).cpu().numpy()[0]
+        img2 = inv_transform(img2).cpu().numpy()[0]
+        # show first image
+        ax = fig.add_subplot(1, 2, 1)
+        plt.imshow(img1[0], cmap=plt.cm.gray)
+        plt.axis("off")
+
+        # show the second image
+        ax = fig.add_subplot(1, 2, 2)
+        plt.imshow(img2[0], cmap=plt.cm.gray)
+        plt.axis("off")
+
+        # show the plot
+        plt.savefig(os.path.join(args.out_path, '{}.png').format(i))
+
     print("Validation: Loss={:.2f}\t Accuracy={:.2f}\t".format(sum(losses)/len(losses), correct / total))

libs/dataset.py

@@ -56,6 +56,8 @@ def create_pairs(self):
         if self.shuffle_pairs:
             np.random.seed(int(time.time()))
             np.random.shuffle(self.indices1)
+        elif self.testing:
+            np.random.seed(int(time.time()))
         else:
             np.random.seed(1)
 
@@ -92,12 +94,12 @@ def __iter__(self):
                 image1 = self.transform(image1).float()
                 image2 = self.transform(image2).float()
 
-            plt.imshow(image1[0])
-            plt.imshow(image2[0])
-            plt.show()
+            # plt.imshow(image1[0])
+            # plt.imshow(image2[0])
+            # plt.show()
 
             if self.testing:
-                yield (image1, image2), torch.FloatTensor([class1==class2]), (image_path1, image_path2)
+                yield (image1, image2), torch.FloatTensor([class1==class2]), (class1, class2)
             else:
                 yield (image1, image2), torch.FloatTensor([class1==class2])
 

siamese/siamese_network.py

@@ -15,15 +15,16 @@ def __init__(self, backbone="resnet18"):
         out_features = list(self.backbone.modules())[-1].out_features
 
         self.cls_head = nn.Sequential(
-            # nn.Dropout(p=0.5),
-            nn.Linear(out_features * 2, 512),
+            nn.Dropout(p=0.5),
+            nn.Linear(out_features, 512),
             nn.BatchNorm1d(512),
-            nn.Sigmoid(),
+            nn.ReLU(),
 
-            # nn.Dropout(p=0.5),
+            nn.Dropout(p=0.5),
             nn.Linear(512, 64),
             nn.BatchNorm1d(64),
             nn.Sigmoid(),
+            nn.Dropout(),
 
             nn.Linear(64, 1),
             nn.Sigmoid(),
@@ -33,7 +34,7 @@ def forward(self, img1, img2):
         feat1 = self.backbone(img1)
         feat2 = self.backbone(img2)
 
-        combined_features = torch.cat((feat1, feat2), dim=-1)
+        combined_features = feat1 * feat2
 
         output = self.cls_head(combined_features)
         return output