class BasicBlock(nn.Module):
def init(self, in_planes, out_planes, dropRate=0.0):
# input dimsnsion을 정하고, output dimension을 정하고(growh_rate임), dropRate를 정함.
super(BasicBlock, self).init()
self.bn1 = nn.BatchNorm2d(in_planes)
self.relu = nn.ReLU(inplace=True) # inplace 하면 input으로 들어온 것 자체를 수정하겠다는 뜻. 메모리 usage가 좀 좋아짐. 하지만 input을 없앰.
self.conv1 = nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=1, padding=1, bias=False)
self.droprate = dropRatedef forward(self, x): out = self.conv1(self.relu(self.bn1(x))) if self.droprate > 0: out = F.dropout(out, p=self.droprate, training=self.training) return torch.cat([x, out], 1)class BottleneckBlock(nn.Module):
def init(self, in_planes, out_planes, dropRate=0.0):
# out_planes => growh_rate를 입력으로 받게 된다.
super(BottleneckBlock, self).init()
inter_planes = out_planes * 4 # bottleneck layer의 conv 1x1 filter chennel 수는 4*growh_rate이다.
self.bn1 = nn.BatchNorm2d(in_planes)
self.relu = nn.ReLU(inplace=True)
self.conv1 = nn.Conv2d(in_planes, inter_planes, kernel_size=1, stride=1, padding=1, bias=False)
self.bn2 = nn.BatchNorm2d(inter_planes)
self.conv2 = nn.Conv2d(inter_planes, out_planes, kernel_size=3, stride=1, padding=0, bias=False)
self.droprate = dropRatedef forward(self, x): out = self.conv1(self.relu(self.bn1(x))) if self.droprate > 0: out = F.dropout(out, p=self.droprate, inplace=False, training=self.training) out = self.conv2(self.relu(self.bn2(out))) if self.droprate > 0: out = F.dropout(out, p=self.droprate, inplace=False, training=self.training) return torch.cat([x, out], 1) # 입력으로 받은 x와 새로 만든 output을 합쳐서 내보낸다class DenseBlock(nn.Module):
def init(self,nb_layers,in_planes,growh_rate,block,dropRate=0.0):
super(DenseBlock,self).init()
self.layer = self._make_layer(block, in_planes, growh_rate, nb_layers, dropRate)
def _make_layer(self,block,in_planes,growh_rate,nb_layers,dropRate):
layers=
for i in range(nb_layers):
layers.append(block(in_planes + i*growh_rate ,growh_rate,dropRate))
return nn.Sequential(*layers)
def forward(self,x):
return self.layer(x)class TransitionBlock(nn.Module):
def init(self,in_planes,out_planes,dropRate=0.0):
super(TransitionBlock,self).init()
self.bn1 = nn.BatchNorm2d(in_planes)
self.relu = nn.ReLU(inplace=True)
self.conv1 = nn.Conv2d(in_planes,out_planes,kernel_size=1,stride=1,padding=0,bias=False)
self.droprate = dropRate
def forward(self,x):
out = self.conv1(self.relu(self.bn1(x)))
if self.droprate>0:
out = F.dropout(out,p=self.droprate,inplace=False,training=self.training)
return F.avg_pool2d(out,2)class DenseNet(nn.Module):
def init(self, depth, num_classes, growh_rate=12, reduction=0.5, bottleneck=True, dropRate=0.0):
super(DenseNet, self).init()
num_of_blocks = 3
in_planes = 16 # 2 * growh_rate
n = (
depth - num_of_blocks - 1) / num_of_blocks # 총 depth에서 첫 conv , 2개의 transit , 마지막 linear 빼고 / num_of_blocks
if reduction != 1:
in_planes = 2 * growh_rate
if bottleneck == True:
in_planes = 2 * growh_rate # 논문에서 Bottleneck + Compression 할 경우 first layer은 2*growh_rate라고 했다.
n = n / 2 # conv 1x1 레이어가 추가되니까 !
block = BottleneckBlock
else:
block = BasicBlockn = int(n) # n = DenseBlock에서 block layer 개수를 의미한다. self.conv1 = nn.Conv2d(3, in_planes, kernel_size=3, stride=1, padding=1, bias=False) # input:RGB -> output:growhR*2 # 1st block # nb_layers,in_planes,growh_rate,block,dropRate self.block1 = DenseBlock(n, in_planes, growh_rate, block, dropRate) in_planes = int(in_planes + n * growh_rate) # 입력 + 레이어 만큼의 growh_rate # in_planes,out_planes,dropRate self.trans1 = TransitionBlock(in_planes, int(math.floor(in_planes * reduction)), dropRate=dropRate) in_planes = int(math.floor(in_planes * reduction)) # 2nd block # nb_layers,in_planes,growh_rate,block,dropRate self.block2 = DenseBlock(n, in_planes, growh_rate, block, dropRate) in_planes = int(in_planes + n * growh_rate) # 입력 + 레이어 만큼의 growh_rate # in_planes,out_planes,dropRate self.trans2 = TransitionBlock(in_planes, int(math.floor(in_planes * reduction)), dropRate=dropRate) in_planes = int(math.floor(in_planes * reduction)) # 3rd block # nb_layers,in_planes,growh_rate,block,dropRate self.block3 = DenseBlock(n, in_planes, growh_rate, block, dropRate) in_planes = int(in_planes + n * growh_rate) # 입력 + 레이어 만큼의 growh_rate self.bn1 = nn.BatchNorm2d(in_planes) self.relu = nn.ReLU(inplace=True) self.fc = nn.Linear(in_planes, num_classes) # 마지막에 ave_pool 후에 1x1 size의 결과만 남음. self.in_planes = in_planes # module 초기화 for m in self.modules(): if isinstance(m, nn.Conv2d): # Conv layer들은 필터에서 나오는 분산 root(2/n)로 normalize 함 # mean = 0 , 분산 = sqrt(2/n) // 이게 무슨 초기화 방법이었는지 기억이 안난다. n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels m.weight.data.normal_(0, math.sqrt(2. / n)) elif isinstance(m, nn.BatchNorm2d): # shifting param이랑 scaling param 초기화(?) m.weight.data.fill_(1) # m.bias.data.zero_() elif isinstance(m, nn.Linear): # linear layer 초기화. m.bias.data.zero_() def forward(self, x): # x : 32*32 out = self.conv1(x) # 32*32 out = self.block1(out) # 32*32 out = self.trans1(out) # 16*16 out = self.block2(out) # 16*16 out = self.trans2(out) # 8*8 out = self.block3(out) # 8*8 out = self.relu(self.bn1(out)) # 8*8 out = F.avg_pool2d(out, 8) # 1*1 out = out.view(-1, self.in_planes) # channel수만 남기 때문에 Linear -> in_planes return self.fc(out)
error
start train
Traceback (most recent call last):
File “/home/jake/Gits/pytorch_/classification/main.py”, line 98, in
train()
File “/home/jake/Gits/pytorch_/classification/main.py”, line 53, in train
outputs = net(inputs)
File “/home/jake/venv/lib/python3.7/site-packages/torch/nn/modules/module.py”, line 541, in call
result = self.forward(input, **kwargs)
File “/home/jake/Gits/pytorch_/classification/models/Densenet.py”, line 150, in forward
out = self.trans1(out) # 1616
File “/home/jake/venv/lib/python3.7/site-packages/torch/nn/modules/module.py”, line 541, in call
result = self.forward(*input, **kwargs)
File “/home/jake/Gits/pytorch_/classification/models/Densenet.py”, line 77, in forward
out = self.conv1(self.relu(self.bn1(x)))
File “/home/jake/venv/lib/python3.7/site-packages/torch/nn/modules/module.py”, line 541, in call
result = self.forward(*input, **kwargs)
File “/home/jake/venv/lib/python3.7/site-packages/torch/nn/modules/batchnorm.py”, line 81, in forward
exponential_average_factor, self.eps)
File “/home/jake/venv/lib/python3.7/site-packages/torch/nn/functional.py”, line 1670, in batch_norm
training, momentum, eps, torch.backends.cudnn.enabled
RuntimeError: running_mean should contain 36 elements not 216
Process finished with exit code 1