def get_conv_output(self, x):
# Layer 2: Convolutional + Batch Norm
conv2_out = self.conv2(x)
bn2_out = self.bn2(conv2_out)
# Layer 3: Convolutional + 2x2 Max Pooling + Batch Norm
conv3_out = self.conv3(bn2_out)
pool3_out = F.max_pool2d(conv3_out, 2)
bn3_out = HF.modified_bn(self.bn3, pool3_out)
# Layer 4: Convolutional + Batch Norm
conv4_out = self.conv4(bn3_out)
bn4_out = HF.modified_bn(self.bn4, conv4_out)
# Build dictionary containing outputs of each layer
conv_out = {
self.CONV2: conv2_out,
self.BN2: bn2_out,
self.CONV3: conv3_out,
self.POOL3: pool3_out,
self.BN3: bn3_out,
self.CONV4: conv4_out,
self.BN4: bn4_out,
}
return conv_out
def get_conv_output(self, x):
# Layer 5: Convolutional + 2x2 Max Pooling + Batch Norm
conv5_out = self.conv5(x)
pool5_out = F.max_pool2d(conv5_out, 2)
bn5_out = self.bn5(pool5_out)
# Layer 6: Convolutional + Batch Norm
conv6_out = self.conv6(bn5_out)
bn6_out = self.bn6(conv6_out)
# Layer 7: Convolutional + 2x2 Max Pooling + Batch Norm
conv7_out = self.conv7(bn6_out)
pool7_out = F.max_pool2d(conv7_out, 2)
bn7_out = HF.modified_bn(self.bn7, pool7_out)
# Layer 8: Convolutional + Batch Norm
conv8_out = self.conv8(bn7_out)
bn8_out = HF.modified_bn(self.bn8, conv8_out)
# Build dictionary containing outputs of each layer
conv_out = {
self.CONV5: conv5_out,
self.POOL5: pool5_out,
self.BN5: bn5_out,
self.CONV6: conv6_out,
self.BN6: bn6_out,
self.CONV7: conv7_out,
self.POOL7: pool7_out,
self.BN7: bn7_out,
self.CONV8: conv8_out,
self.BN8: bn8_out,
}
return conv_out
def get_conv_output(self, x):
# Layer 4: Convolutional + Batch Norm
conv4_out = self.conv4(x)
bn4_out = HF.modified_bn(self.bn4, conv4_out)
# Build dictionary containing outputs of each layer
conv_out = {
self.CONV4: conv4_out,
self.BN4: bn4_out,
}
return conv_out
def get_conv_output(self, x):
# Layer 8: Convolutional + Batch Norm
conv8_out = self.conv8(x)
bn8_out = HF.modified_bn(self.bn8, conv8_out)
# Build dictionary containing outputs of each layer
conv_out = {
self.CONV8: conv8_out,
self.BN8: bn8_out,
}
return conv_out
def forward(self, x):
# Compute the output feature map from the convolutional layers
out = self.get_conv_output(x)
# Layer 9: FC + Batch Norm
fc9_out = self.fc9(out[self.CONV_OUTPUT])
bn9_out = HF.modified_bn(self.bn9, fc9_out)
# Linear FC layer, outputs are the class scores
fc10_out = self.fc10(bn9_out).view(-1, self.NUM_CLASSES)
# Build dictionary containing outputs from convolutional and FC layers
out[self.FC9] = fc9_out
out[self.BN9] = bn9_out
out[self.FC10] = fc10_out
return out
def forward(self, x):
out = {}
# Hidden Layer: FC + Batch Norm
fc1_out = self.fc1(x if len(self.get_input_shape()) >= 3 else x.
view(x.size(0), x.size(1), 1, 1))
bn1_out = HF.modified_bn(self.bn1, fc1_out)
# Output Layer, outputs are the class scores
fc2_out = self.fc2(bn1_out).view(-1, self.NUM_CLASSES)
# Build dictionary containing outputs from convolutional and FC layers
out[self.FC1] = fc1_out
out[self.BN1] = bn1_out
out[self.FC2] = fc2_out
return out
def forward(self, x):
# Compute the output feature map from the convolutional layers
out = self.get_conv_output(x)
# Layer 5: FC + Batch Norm
fc5_out = self.fc5(out[self.CONV_OUTPUT])
bn5_out = HF.modified_bn(self.bn5, fc5_out)
# Linear FC layer, outputs are the class scores
fc6_out = self.fc6(bn5_out).view(-1, self.NUM_CLASSES)
# Build dictionary containing outputs from convolutional and FC layers
out[self.FC5] = fc5_out
out[self.BN5] = bn5_out
out[self.FC6] = fc6_out
out[self.CLASS_SCORES] = {P.KEY_CLASS_SCORES: fc6_out}
return out
def get_conv_output(self, x):
# Layer 1: Convolutional + 3x3 Max Pooling + Batch Norm
conv1_out = self.conv1(x)
pool1_out = F.max_pool2d(conv1_out, 3)
bn1_out = self.bn1(pool1_out)
# Layer 2: Convolutional + Batch Norm
conv2_out = self.conv2(bn1_out)
bn2_out = self.bn2(conv2_out)
# Layer 3: Convolutional + 2x2 Max Pooling + Batch Norm
conv3_out = self.conv3(bn2_out)
pool3_out = F.max_pool2d(conv3_out, 2)
bn3_out = self.bn3(pool3_out)
# Layer 4: Convolutional + Batch Norm
conv4_out = self.conv4(bn3_out)
bn4_out = self.bn4(conv4_out)
# Layer 5: Convolutional + 2x2 Max Pooling + Batch Norm
conv5_out = self.conv5(bn4_out)
pool5_out = F.max_pool2d(conv5_out, 2)
bn5_out = self.bn5(pool5_out)
# Layer 6: Convolutional + Batch Norm
conv6_out = self.conv6(bn5_out)
bn6_out = self.bn6(conv6_out)
# Layer 7: Convolutional + 2x2 Max Pooling + Batch Norm
conv7_out = self.conv7(bn6_out)
pool7_out = F.max_pool2d(conv7_out, 2)
bn7_out = HF.modified_bn(self.bn7, pool7_out)
# Layer 8: Convolutional + Batch Norm
conv8_out = self.conv8(bn7_out)
bn8_out = HF.modified_bn(self.bn8, conv8_out)
# Build dictionary containing outputs of each layer
conv_out = {
self.CONV1: conv1_out,
self.POOL1: pool1_out,
self.BN1: bn1_out,
self.CONV2: conv2_out,
self.BN2: bn2_out,
self.CONV3: conv3_out,
self.POOL3: pool3_out,
self.BN3: bn3_out,
self.CONV4: conv4_out,
self.BN4: bn4_out,
self.CONV5: conv5_out,
self.POOL5: pool5_out,
self.BN5: bn5_out,
self.CONV6: conv6_out,
self.BN6: bn6_out,
self.CONV7: conv7_out,
self.POOL7: pool7_out,
self.BN7: bn7_out,
self.CONV8: conv8_out,
self.BN8: bn8_out,
}
return conv_out
