Added dummy input for one-shot size-checking

checkpoint.py

@@ -0,0 +1,8 @@
+import torch
+import shutil
+
+def save_checkpoint(state, is_best: bool, task_id, fname="checkpoint.pth.tar"):
+    fdir = "./"+str(task_id)+"/"
+    torch.save(state, fdir + fname)
+    if is_best:
+        shutil.copyfile(fdir + fname, fdir + "best.pth.tar")

model/stn.py

@@ -18,8 +18,13 @@ class STNet(nn.Module):
         assert 5 > len(input_size) > 2  # single or batch ([N, ]C, H, W)
         if len(input_size) == 3:
             channels, height, width = input_size
+            _dummy_size_ = torch.Size([1]) + input_size
         else:
             channels, height, width = input_size[1:]
+            _dummy_size_ = input_size
+
+        # shape checking
+        _dummy_x_ = torch.zeros(_dummy_size_)
 
         # (3.1) Spatial transformer localization-network
         self.localization_net = nn.Sequential(      # (N, C, H, W)
@@ -30,11 +35,8 @@ class STNet(nn.Module):
             nn.MaxPool2d(2, stride=2),              # (N, 10, (((H-6)/2)-4)/2, (((H-6)/2)-4)/2)
             nn.ReLU(True)
         ) # -> (N, 10, (((H-6)/2)-4)/2, (((H-6)/2)-4)/2)
-        self._loc_net_out_shape = (
-            10,
-            (((height - 6) // 2) - 4) // 2,
-            (((width - 6) // 2) - 4) // 2
-        ) # TODO: PLEASE let me know if there are better ways of doing this...
+        _dummy_x_ = self.localization_net(_dummy_x_)
+        self._loc_net_out_shape = _dummy_x_.shape[1:]
 
         # (3.2) Regressor for the 3 * 2 affine matrix
         self.fc_loc = nn.Sequential(
@@ -53,6 +55,10 @@ class STNet(nn.Module):
                 dtype=torch.float
             )
         )
+        _dummy_x_ = self.fc_loc(
+            _dummy_x_.view(-1, np.prod(self._loc_net_out_shape))
+        )
+        return
 
 
     # Spatial transformer network forward function
@@ -205,8 +211,8 @@ if __name__ == "__main__":
             axarr[1].imshow(out_grid)
             axarr[1].set_title('Transformed Images')
 
-    for epoch in range(1, 20 + 1):
-        train(epoch)
+    for epoch in range(10):
+        train(epoch + 1)
         valid()
 
     # Visualize the STN transformation on some input batch