Yes, with sum you are right: I take vgg_a and vgg_b and I sum layer per layer their weights.
For showing the parameters, I mean this, I am not doing in a right way the print of the weights.
I have these:
model1 = VGG16((1,32,32),batch_norm=True)
model2 = VGG16((1,32,32),batch_norm=True)
model3 = VGG16((1,32,32),batch_norm=True)
classifier = Classifier(num_classes=1)
nets = [model1,model2,model3,classifier]
where VGG16 and Classifier are:
class VGGBlock(nn.Module):
def __init__(self, in_channels, out_channels,batch_norm=False):
super().__init__()
conv2_params = {'kernel_size': (3, 3),
'stride' : (1, 1),
'padding' : 1
}
noop = lambda x : x
self._batch_norm = batch_norm
self.conv1 = nn.Conv2d(in_channels=in_channels,out_channels=out_channels , **conv2_params)
self.bn1 = nn.BatchNorm2d(out_channels) if batch_norm else noop
self.conv2 = nn.Conv2d(in_channels=out_channels,out_channels=out_channels, **conv2_params)
self.bn2 = nn.BatchNorm2d(out_channels) if batch_norm else noop
self.max_pooling = nn.MaxPool2d(kernel_size=(2, 2), stride=(2, 2))
@property
def batch_norm(self):
return self._batch_norm
def forward(self,x):
x = self.conv1(x)
x = self.bn1(x)
x = F.relu(x)
x = self.conv2(x)
x = self.bn2(x)
x = F.relu(x)
x = self.max_pooling(x)
return x
class VGG16(nn.Module):
def __init__(self, input_size, batch_norm=False):
super(VGG16, self).__init__()
self.in_channels,self.in_width,self.in_height = input_size
self.block_1 = VGGBlock(self.in_channels,64,batch_norm=batch_norm)
self.block_2 = VGGBlock(64, 128,batch_norm=batch_norm)
self.block_3 = VGGBlock(128, 256,batch_norm=batch_norm)
self.block_4 = VGGBlock(256,512,batch_norm=batch_norm)
@property
def input_size(self):
return self.in_channels,self.in_width,self.in_height
def forward(self, x):
x = self.block_1(x)
x = self.block_2(x)
x = self.block_3(x)
x = self.block_4(x)
# x = self.avgpool(x)
x = torch.flatten(x,1)
return x
class Classifier(nn.Module):
def __init__(self,num_classes=1):
super().__init__()
self.classifier = nn.Sequential(
nn.Linear(2048, 2048),
nn.ReLU(True),
nn.Dropout(p=0.5),
nn.Linear(2048, 512),
nn.ReLU(True),
nn.Dropout(p=0.5),
nn.Linear(512, num_classes)
)
def forward(self,x):
return self.classifier(x)
And in the train there is this part to do the sum:
for n in model_a:
layer_a = model_a[n]
layer_b = model_b[n]
layer_c = model_c[n]
if (isinstance(layer_a,nn.Conv2d)):
loss += lambda_reg * reg_loss(combo_fn(layer_a.weight,layer_b.weight),layer_c.weight)
if (layer_a.bias is not None):
loss += lambda_reg * reg_loss(combo_fn(layer_a.bias, layer_b.bias), layer_c.bias)