evaluation of different pooling types

python/main.py

@@ -38,6 +38,20 @@ SPACE = "RGB"
 
 LEARNING_RATE = 0.001 # DEAFULT IS 0.001
 
+##################################
+##       POOLING METHODS        ##
+##################################
+# Here we consistently use RGB color space and LEARNING_RATE = 0.001
+#
+# Max pooling: (73% goalnet)
+# Avg pooling: (59% goalnet)
+# Max + Avg + Max (72% goalnet)
+# Avg + Max + Avg (62 % goalnet)
+#
+# Max pooling, 8x8,4x4,2x2: (73% goalnet)
+# Max pooling, 16x16,9x9,2x2 (68% goalnet)
+# Max pooling, 24x24,4x4,2x2 (65% goalnet)
+
 ##################################
 ##        PREPARE DATA          ##
 ##################################
@@ -95,7 +109,7 @@ objective = nn.CrossEntropyLoss()
 optimizer = torch.optim.Adam(model.parameters(), lr=LEARNING_RATE)
 
 ##################################
-##         TRAN NETWORK         ##
+##         TRAIN NETWORK         ##
 ##################################
 
 start_epoch = 1

python/models/goalNet.py

@@ -16,15 +16,24 @@ class goalNet(nn.Module):
             # Input rgb channels, output 32 channels (32 parallel computations), kernel size 5x5
             nn.Conv2d(channels, 32, 5, padding=2),
             # outputs 16x16xc featurmap. Stride 's' downscales by a factor 's'
-            nn.MaxPool2d(kernel_size=8, padding=4, stride=2),
+            #nn.MaxPool2d(kernel_size=24, padding=11, stride=2),
+            #nn.MaxPool2d(kernel_size=16, padding=7, stride=2),
+            nn.MaxPool2d(kernel_size=8, padding=4, stride=2),            
+            #nn.AvgPool2d(kernel_size=8, padding=4, stride=2),
             nn.ReLU())
 
         self.seq2 = nn.Sequential(
             nn.Conv2d(32, 32, 5, padding=2), nn.ReLU(),
-            nn.MaxPool2d(kernel_size=4, padding=2, stride=2))
+            #nn.MaxPool2d(kernel_size=4, padding=1, stride=2)
+            #nn.MaxPool2d(kernel_size=9, padding=0, stride=1)
+            nn.MaxPool2d(kernel_size=4, padding=2, stride=2)
+            #nn.AvgPool2d(kernel_size=4, padding=2, stride=2)
+            )
 
         self.seq3 = nn.Sequential(nn.Conv2d(32, 64, 5, padding=2), nn.ReLU(),
-                                  nn.MaxPool2d(kernel_size=2, stride=2))
+                                  nn.MaxPool2d(kernel_size=2, stride=2)
+                                  #nn.AvgPool2d(kernel_size=2, stride=2)
+                                 )
 
         # Fully connected since expects a 4x4xc feature map
         self.fc1 = nn.Conv2d(64, 64, 4)