pytorch实现resnet18(34,101,152)、vgg16对cifar10进行分类
1、pytorch简介
pytorch是由facebook所开源的深度学习框架，其框架重于强调于动态流图建立，其不同于google的tensorflow其静态Graph概念，其各有千秋。tensorflow更多贡献于distribution sysytem，企业分布式训练体系架构非常适合。pytorch是后来居上，其构造的框架API也在不断改善，但是pytorch对于类的构造、继承等发挥淋漓尽致，方便程序员、学者非常容易构建训练模型和动态Gragh。TensorflowGPU训练会先占用训练GPU上的剩余空间，即使它未发生使用，这也是资源浪费行为，在于Torch而是实际使用显存占位，这样使用实际有利于共享资源的深度学习爱好者或是其他学习者----这也可能是pytorch在高校、公司实验室受欢迎的原因之一。
2、数据与数据预处理
原来的Carf10是来自Imagenet的10个分类，而且其数据进行压缩至32*32的像素，一般用于算法高校性的测试和处理。这里我使用的AlexNet、VGG、Resnet18、Resnet34、Resnet101、Resnet152进行处理训练，所以前期需要进行数据的与处理：
transform = transforms.Compose([transforms.Resize((240,240)),
transforms.RandomCrop((224,224)),transforms.ToTensor(),
transforms.Normalize([0.485,0.456,0.406],[0.229,0.224,0.225])])
#First parameter: Resize the size of Image
#Second : RandomCrop is to increase Data Augment
#Third : ToTensor means it convert the type of numpy data into the tensor of pytorch
# Forth: Normalize Data数据提取自己封装了一个数据提取的类：
class GetTestData():
def __init__(self, loadfile , batch_size , num_work ,trainform,mode='carf10'):
self.load_files = loadfile
self.batch_size = batch_size
self.num_work = num_work # using multi-thread methods to read Datasets
self.mode = mode # tags that decide to choose different dataset
self.transform =trainform
##
##
def GetRawData(self,index=1):
if self.mode == 'carf10' :
train_data = datasets.CIFAR10(self.load_files,train=True,transform=self.transform.ToTensor())
test_data = datasets.CIFAR10(self.load_files,train=False,transform=self.transform.ToTensor())
elif self.mode == 'minist':
train_data = datasets.MNIST(self.load_files,train=True,transform=self.transform.ToTensor())
test_data = datasets.MNIST(self.load_files,train=False,transform=self.transform.ToTensor())
else :
train_data = datasets.FashionMNIST(self.load_files,train=True,transform=self.transform.ToTensor())
test_data = datasets.FashionMNIST(self.load_files,train=False,transform=self.transform.ToTensor())
if index == 1:
train_dataset = Subset(train_data,train_index)
valid_dataset = Subset(train_data,valid_index)
train_loader = DataLoader(train_data,batch_size=self.batch_size,shuffle=True,num_workers=self.num_work)
valid_loader = DataLoader(valid_dataset,batch_size=self.batch_size,shuffle=False,num_workers=self.num_work)
test_loader = DataLoader(test_data,batch_size=self.batch_size,shuffle=False,num_workers=self.num_work)
return train_loader , valid_loader , test_loader
else :
train_loader = DataLoader(train_data,batch_size=self.batch_size,shuffle=True,num_workers=self.num_work)
test_loader = DataLoader(test_data,batch_size=self.batch_size,shuffle=False,num_workers=self.num_work)
return train_loader , test_loader
def GetResize(self,index = 1):
if self.mode == 'carf10' :
train_data = datasets.CIFAR10(self.load_files,train=True,transform=self.transform)
test_data = datasets.CIFAR10(self.load_files,train=False,transform=self.transform)
elif self.mode == 'minist':
train_data = datasets.MNIST(self.load_files,train=True,transform=self.transform)
test_data = datasets.MNIST(self.load_files,train=False,transform=self.transform )
else :
train_data = datasets.FashionMNIST(self.load_files,train=True,transform=self.transform )
test_data = datasets.FashionMNIST(self.load_files,train=False,transform=self.transform)
if index == 1:
train_dataset = Subset(train_data,train_index)
valid_dataset = Subset(train_data,valid_index)
train_loader = DataLoader(train_data,batch_size=self.batch_size,shuffle=True,num_workers=self.num_work)
valid_loader = DataLoader(valid_dataset,batch_size=self.batch_size,shuffle=False,num_workers=self.num_work)
test_loader = DataLoader(test_data,batch_size=self.batch_size,shuffle=False,num_workers=self.num_work)
return train_loader , valid_loader , test_loader
else :
train_loader = DataLoader(train_data,batch_size=self.batch_size,shuffle=True,num_workers=self.num_work)
test_loader = DataLoader(test_data,batch_size=self.batch_size,shuffle=False,num_workers=self.num_work)
return train_loader , test_loader
这里我之前作数据处理是发生了一个常见的错误“
ERROR Warning TypeError: tensor is not a torch image.
# 这里我是把tensor放在Normaliztion后面，提示出错，只需要将其改在Normalization前面，这是API先将其转换为tensor再进行减去RGB均值
”
3、网络结构实现
可以清晰看到resnet18、34、50、101、152的不同结构，这样我设计两个部分用于调用：
##### resnet 18 34
#[2,2,2,2]
class BasicBlock(nn.Module):
expansion = 1
def __init__(self, inplanes, planes, stride=1, downsample=None):
super(BasicBlock, self).__init__()
self.conv1 = conv3x3(inplanes, planes, stride)
self.bn1 = nn.BatchNorm2d(planes)
self.relu = nn.ReLU(inplace=True)
self.conv2 = conv3x3(planes, planes)
self.bn2 = nn.BatchNorm2d(planes)
self.downsample = downsample
self.stride = stride
def forward(self, x):
residual = x
out = self.conv1(x)
out = self.bn1(out)
out = self.relu(out)
out = self.conv2(out)
out = self.bn2(out)
if self.downsample is not None:
residual = self.downsample(x)
out += residual
out = self.relu(out)
return out
class ResNet(nn.Module):
def __init__(self, block, layers, num_classes, grayscale=False):
self.inplanes = 64
if grayscale:
in_dim = 1
else:
in_dim = 3
super(ResNet, self).__init__()
self.conv1 = nn.Conv2d(in_dim, 64, kernel_size=7, stride=2, padding=3,
bias=False)
self.bn1 = nn.BatchNorm2d(64)
self.relu = nn.ReLU(inplace=True)
self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
self.layer1 = self._make_layer(block, 64, layers[0])
self.layer2 = self._make_layer(block, 128, layers[1], stride=2)
self.layer3 = self._make_layer(block, 256, layers[2], stride=2)
self.layer4 = self._make_layer(block, 512, layers[3], stride=2)
self.avgpool = nn.AvgPool2d(7, stride=1)
self.fc = nn.Linear(512 * block.expansion, num_classes)
for m in self.modules():
if isinstance(m, nn.Conv2d):
n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
m.weight.data.normal_(0, (2. / n)**.5)
elif isinstance(m, nn.BatchNorm2d):
m.weight.data.fill_(1)
m.bias.data.zero_()
def _make_layer(self, block, planes, blocks, stride=1):
downsample = None
if stride != 1 or self.inplanes != planes * block.expansion:
downsample = nn.Sequential(
nn.Conv2d(self.inplanes, planes * block.expansion,
kernel_size=1, stride=stride, bias=False),
nn.BatchNorm2d(planes * block.expansion),
)
layers = []
layers.append(block(self.inplanes, planes, stride, downsample))
self.inplanes = planes * block.expansion
for i in range(1, blocks):
layers.append(block(self.inplanes, planes))
return nn.Sequential(*layers)
def forward(self, x):
x = self.conv1(x)
x = self.bn1(x)
x = self.relu(x)
x = self.maxpool(x)
x = self.layer1(x)
x = self.layer2(x)
x = self.layer3(x)
x = self.layer4(x)
x = self.avgpool(x)
x = x.view(x.size(0), -1)
logits = self.fc(x)
probas = F.softmax(logits, dim=1)
return logits, probasdef conv3x3(inplane , outplane, stride=1):
return nn.Conv2d(inplane,outplane,kernel_size=3,stride=stride,padding=1)
## ZhiwenXiao
class Bottleneck(nn.Module):
expansion = 4
def __init__(self, inplanes, planes, stride=1, downsample=None):
super(Bottleneck, self).__init__()
self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=1, bias=False)
self.bn1 = nn.BatchNorm2d(planes)
self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride,
padding=1, bias=False)
self.bn2 = nn.BatchNorm2d(planes)
self.conv3 = nn.Conv2d(planes, planes * 4, kernel_size=1, bias=False)
self.bn3 = nn.BatchNorm2d(planes * 4)
self.relu = nn.ReLU(inplace=True)
self.downsample = downsample
self.stride = stride
def forward(self, x):
residual = x
out = self.conv1(x)
out = self.bn1(out)
out = self.relu(out)
out = self.conv2(out)
out = self.bn2(out)
out = self.relu(out)
out = self.conv3(out)
out = self.bn3(out)
if self.downsample is not None:
residual = self.downsample(x)
out += residual
out = self.relu(out)
return out
class ResNet101(nn.Module):
def __init__(self, block, layers, num_classes, grayscale):
self.inplanes = 64
if grayscale:
in_dim = 1
else:
in_dim = 3
super(ResNet101, self).__init__()
self.conv1 = nn.Conv2d(in_dim, 64, kernel_size=7, stride=2, padding=3,
bias=False)
self.bn1 = nn.BatchNorm2d(64)
self.relu = nn.ReLU(inplace=True)
self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
self.layer1 = self._make_layer(block, 64, layers[0])
self.layer2 = self._make_layer(block, 128, layers[1], stride=2)
self.layer3 = self._make_layer(block, 256, layers[2], stride=2)
self.layer4 = self._make_layer(block, 512, layers[3], stride=2)
self.avgpool = nn.AvgPool2d(7, stride=1, padding=0)
#self.fc = nn.Linear(2048 * block.expansion, num_classes)
self.fc = nn.Linear(2048, num_classes)
for m in self.modules():
if isinstance(m, nn.Conv2d):
n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
m.weight.data.normal_(0, (2. / n)**.5)
elif isinstance(m, nn.BatchNorm2d):
m.weight.data.fill_(1)
m.bias.data.zero_()
def _make_layer(self, block, planes, blocks, stride=1):
downsample = None
if stride != 1 or self.inplanes != planes * block.expansion:
downsample = nn.Sequential(
nn.Conv2d(self.inplanes, planes * block.expansion,
kernel_size=1, stride=stride, bias=False),
nn.BatchNorm2d(planes * block.expansion),
)
layers = []
layers.append(block(self.inplanes, planes, stride, downsample))
self.inplanes = planes * block.expansion
for i in range(1, blocks):
layers.append(block(self.inplanes, planes))
return nn.Sequential(*layers)
def forward(self, x):
x = self.conv1(x)
x = self.bn1(x)
x = self.relu(x)
x = self.maxpool(x)
x = self.layer1(x)
x = self.layer2(x)
x = self.layer3(x)
x = self.layer4(x)
x = self.avgpool(x)
x = x.view(x.size(0), -1)
logits = self.fc(x)
probas = F.softmax(logits, dim=1)
return logits, probasVGG16结构网络为：
class VGG16(nn.Module):
def __init__(self,num_classes,grayscale=False):
dim = None
if grayscale==True:
dim = 1
else :
dim = 3
super(VGG16,self).__init__()
self.vgg_bone = nn.Sequential(
nn.Conv2d(dim,64,kernel_size=3,padding=1),
nn.ReLU(inplace=True),
nn.Conv2d(64,64,kernel_size=3,padding=1),
nn.ReLU(inplace=True),
nn.MaxPool2d(kernel_size=2,stride=2,padding=0),
nn.Conv2d(64,128,kernel_size=3,stride=1,padding=1),
nn.ReLU(inplace=True),
nn.Conv2d(128,128,kernel_size=3, padding=1 ),
nn.ReLU(inplace=True),
nn.MaxPool2d(kernel_size=2,stride=2,padding=0),
nn.Conv2d(128,256,kernel_size=3,stride=1,padding=1),
nn.ReLU(inplace=True),
nn.Conv2d(256,256,kernel_size=3,stride=1,padding=1),
nn.ReLU(inplace=True),
nn.Conv2d(256,256,kernel_size=3,stride=1,padding=1),
nn.ReLU(inplace=True),
nn.MaxPool2d(kernel_size=2,stride=2,padding=0),
nn.Conv2d(256,512,kernel_size=3,stride=1,padding=1),
nn.ReLU(inplace=True),
nn.Conv2d(512,512,kernel_size=3,stride=1,padding=1),
nn.ReLU(inplace=True),
nn.Conv2d(512,512,kernel_size=3,stride=1,padding=1),
nn.ReLU(inplace=True),
nn.MaxPool2d(kernel_size=2,stride=2,padding=0),
nn.Conv2d(512,512,kernel_size=3,stride=1,padding=1),
nn.ReLU(inplace=True),
nn.Conv2d(512,512,kernel_size=3,stride=1,padding=1),
nn.ReLU(inplace=True),
nn.Conv2d(512,512,kernel_size=3,stride=1,padding=1),
nn.ReLU(inplace=True),
nn.MaxPool2d(kernel_size=2,stride=2,padding=0),
)
self.vgg_logit = nn.Sequential(
nn.Linear(7*7*512,4096),
nn.ReLU(inplace=True),
nn.Dropout(p=0.5),
nn.Linear(4096,4096),
nn.ReLU(inplace=True),
nn.Dropout(p=0.5),
nn.Linear(4096,num_classes),
)
for m in self.modules():
if isinstance(m ,nn.Conv2d ):
m.weight.detach().normal_(0,0.05)
if m.bias is not None :
m.bias.data.detach().zero_()
elif isinstance(m,nn.Linear):
m.weight.detach().normal_(0,0.05)
m.bias.detach().detach().zero_()
def forward(self,x):
x =self.vgg_bone(x)
x =x.view(x.size(0),-1)
logit = self.vgg_logit(x)
prob = F.softmax(logit)
return logit,prob4、训练与对比
训练参数较大，设置的迭代的epoch设置的不是特别大
当时标题问题，这是ResNet18
这是resnet50
resnet101 with 128 batch_size only trains 10 epochs