PyTorch 迁移学习 (Transfer Learning) 代码详解
2024-05-22 21:00:00
PyTorch 迁移学习 代码详解
- 概述
- 为什么使用迁移学习
- 更好的结果
- 节省时间
- 加载模型
- ResNet152
- 冻层实现
- 模型初始化
- 获取需更新参数
- 训练模型
- 获取数据
- 完整代码
概述
迁移学习 (Transfer Learning) 是把已学训练好的模型参数用作新训练模型的起始参数. 迁移学习是深度学习中非常重要和常用的一个策略.
为什么使用迁移学习
更好的结果
迁移学习 (Transfer Learning) 可以帮助我们得到更好的结果.
当我们手上的数据比较少的时候, 训练非常容易造成过拟合的现象. 使用迁移学习可以帮助我们通过更少的训练数据达到更好的效果. 使得模型的泛化能力更强, 训练过程更稳定.
节省时间
迁移学习 (Transfer Learning) 可以帮助我们节省时间.
通过迁徙学习, 我们站在了巨人的肩膀上. 利用前人花大量时间训练好的参数, 能帮助我们在模型的训练上节省大把的时间.
加载模型
首先我们需要加载模型, 并指定层数. 常用的模型有:
- VGG
- ResNet
- SqueezeNet
- DenseNet
- Inception
- GoogLeNet
- ShuffleNet
- MobileNet
官网 API
ResNet152
我们将使用 ResNet 152 和 CIFAR 100 来举例.
冻层实现
def set_parameter_requires_grad(model, feature_extracting):"""是否保留梯度, 实现冻层:param model: 模型:param feature_extracting: 是否冻层:return: 无返回值"""if feature_extracting: # 如果冻层for param in model.parameters(): # 遍历每个权重参数param.requires_grad = False # 保留梯度为False
模型初始化
def initialize_model(model_name, num_classes, feature_exact, use_pretrained=True):"""初始化模型:param model_name: 模型名字:param num_classes: 类别数:param feature_exact: 是否冻层:param use_pretrained: 是否下载模型:return: 返回模型,"""model_ft = Noneif model_name == "resnet":"""Resnet152"""# 加载模型model_ft = models.resnet152(pretrained=use_pretrained) # 下载参数set_parameter_requires_grad(model_ft, feature_exact) # 冻层# 修改全连接层num_features = model_ft.fc.in_featuresmodel_ft.fc = torch.nn.Sequential(torch.nn.Linear(num_features, num_classes),torch.nn.LogSoftmax(dim=1))# 返回初始化好的模型return model_ft
获取需更新参数
def parameter_to_update(model):"""获取需要更新的参数:param model: 模型:return: 需要更新的参数列表"""print("Params to learn")param_array = model.parameters()if feature_exact:param_array = []for name, param, in model.named_parameters():if param.requires_grad == True:param_array.append(param)print("\t", name)else:for name, param, in model.named_parameters():if param.requires_grad == True:print("\t", name)return param_array
训练模型
def train_model(model, dataloaders, citerion, optimizer, filename, num_epochs=25):# 获取起始时间since = time.time()# 初始化参数best_acc = 0val_acc_history = []train_acc_history = []train_losses = []valid_losses = []LRs = [optimizer.param_groups[0]["lr"]]best_model_weights = copy.deepcopy(model.state_dict())for epoch in range(num_epochs):print("Epoch {}/{}".format(epoch, num_epochs - 1))print("-" * 10)# 训练和验证for phase in ["train", "valid"]:if phase == "train":model.train() # 训练else:model.eval() # 验证running_loss = 0.0running_corrects = 0# 遍历数据for inputs, labels in dataloaders[phase]:inputs = inputs.to(device)labels = labels.to(device)# 梯度清零optimizer.zero_grad()# 只有训练的时候计算和更新梯度with torch.set_grad_enabled(phase == "train"):outputs = model(inputs)_, preds = torch.max(outputs, 1)# 计算损失loss = criterion(outputs, labels)# 训练阶段更新权重if phase == "train":loss.backward()optimizer.step()# 计算损失running_loss += loss.item() * inputs.size(0)running_corrects += torch.sum(preds == labels.data)epoch_loss = running_loss / len(dataloaders[phase].dataset)epoch_acc = running_corrects.double() / len(dataloaders[phase].dataset)time_eplased = time.time() - sinceprint("Time elapsed {:.0f}m {:.0f}s".format(time_eplased // 60, time_eplased % 60))print("{} Loss: {:.4f} Acc: {:.4f}".format(phase, epoch_loss, epoch_acc))# 得到最好的模型if phase == "valid" and epoch_acc > best_acc:best_acc = epoch_accbest_model_weights = copy.deepcopy(model.state_dict())state = {"state_dict": model.state_dict(),"best_acc": best_acc,"optimizer": optimizer.state_dict(),}torch.save(state, filename)if phase == "valid":val_acc_history.append(epoch_acc)valid_losses.append(epoch_loss)scheduler.step(epoch_loss)if phase == "train":train_acc_history.append(epoch_acc)train_losses.append(epoch_loss)print("Optimizer learning rate: {:.7f}".format(optimizer.param_groups[0]["lr"]))LRs.append(optimizer.param_groups[0]["lr"])print()time_eplased = time.time() - sinceprint("Training complete in {:.0f}m {:.0f}s".format(time_eplased // 60, time_eplased % 60))print("Best val Acc: {:4f}".format(best_acc))# 训练完后用最好的一次当做模型最终的结果model.load_state_dict(best_model_weights)# 返回return model, val_acc_history, train_acc_history, valid_losses, train_losses, LRs
获取数据
def get_data():"""获取数据"""# 获取测试集train = torchvision.datasets.CIFAR100(root="./mnt", train=True, download=True,transform=torchvision.transforms.Compose([torchvision.transforms.ToTensor(), # 转换成张量torchvision.transforms.Normalize((0.1307,), (0.3081,)) # 标准化]))train_loader = DataLoader(train, batch_size=batch_size) # 分割测试集# 获取测试集test = torchvision.datasets.CIFAR100(root="./mnt", train=False, download=True,transform=torchvision.transforms.Compose([torchvision.transforms.ToTensor(), # 转换成张量torchvision.transforms.Normalize((0.1307,), (0.3081,)) # 标准化]))test_loader = DataLoader(test, batch_size=batch_size) # 分割训练data_loader = {"train": train_loader, "valid": test_loader}# 返回分割好的训练集和测试集return data_loader
完整代码
完整代码:
import copy
import torch
from torch.utils.data import DataLoader
import time
from torchsummary import summary
import torchvision
import torchvision.models as modelsdef set_parameter_requires_grad(model, feature_extracting):"""是否保留梯度, 实现冻层:param model: 模型:param feature_extracting: 是否冻层:return: 无返回值"""if feature_extracting: # 如果冻层for param in model.parameters(): # 遍历每个权重参数param.requires_grad = False # 保留梯度为Falsedef initialize_model(model_name, num_classes, feature_exact, use_pretrained=True):"""初始化模型:param model_name: 模型名字:param num_classes: 类别数:param feature_exact: 是否冻层:param use_pretrained: 是否下载模型:return: 返回模型,"""model_ft = Noneif model_name == "resnet":"""Resnet152"""# 加载模型model_ft = models.resnet152(pretrained=use_pretrained) # 下载参数set_parameter_requires_grad(model_ft, feature_exact) # 冻层# 修改全连接层num_features = model_ft.fc.in_featuresmodel_ft.fc = torch.nn.Sequential(torch.nn.Linear(num_features, num_classes),torch.nn.LogSoftmax(dim=1))# 返回初始化好的模型return model_ftdef parameter_to_update(model):"""获取需要更新的参数:param model: 模型:return: 需要更新的参数列表"""print("Params to learn")param_array = model.parameters()if feature_exact:param_array = []for name, param, in model.named_parameters():if param.requires_grad == True:param_array.append(param)print("\t", name)else:for name, param, in model.named_parameters():if param.requires_grad == True:print("\t", name)return param_arraydef train_model(model, dataloaders, citerion, optimizer, filename, num_epochs=25):# 获取起始时间since = time.time()# 初始化参数best_acc = 0val_acc_history = []train_acc_history = []train_losses = []valid_losses = []LRs = [optimizer.param_groups[0]["lr"]]best_model_weights = copy.deepcopy(model.state_dict())for epoch in range(num_epochs):print("Epoch {}/{}".format(epoch, num_epochs - 1))print("-" * 10)# 训练和验证for phase in ["train", "valid"]:if phase == "train":model.train() # 训练else:model.eval() # 验证running_loss = 0.0running_corrects = 0# 遍历数据for inputs, labels in dataloaders[phase]:inputs = inputs.to(device)labels = labels.to(device)# 梯度清零optimizer.zero_grad()# 只有训练的时候计算和更新梯度with torch.set_grad_enabled(phase == "train"):outputs = model(inputs)_, preds = torch.max(outputs, 1)# 计算损失loss = criterion(outputs, labels)# 训练阶段更新权重if phase == "train":loss.backward()optimizer.step()# 计算损失running_loss += loss.item() * inputs.size(0)running_corrects += torch.sum(preds == labels.data)epoch_loss = running_loss / len(dataloaders[phase].dataset)epoch_acc = running_corrects.double() / len(dataloaders[phase].dataset)time_eplased = time.time() - sinceprint("Time elapsed {:.0f}m {:.0f}s".format(time_eplased // 60, time_eplased % 60))print("{} Loss: {:.4f} Acc: {:.4f}".format(phase, epoch_loss, epoch_acc))# 得到最好的模型if phase == "valid" and epoch_acc > best_acc:best_acc = epoch_accbest_model_weights = copy.deepcopy(model.state_dict())state = {"state_dict": model.state_dict(),"best_acc": best_acc,"optimizer": optimizer.state_dict(),}torch.save(state, filename)if phase == "valid":val_acc_history.append(epoch_acc)valid_losses.append(epoch_loss)scheduler.step(epoch_loss)if phase == "train":train_acc_history.append(epoch_acc)train_losses.append(epoch_loss)print("Optimizer learning rate: {:.7f}".format(optimizer.param_groups[0]["lr"]))LRs.append(optimizer.param_groups[0]["lr"])print()time_eplased = time.time() - sinceprint("Training complete in {:.0f}m {:.0f}s".format(time_eplased // 60, time_eplased % 60))print("Best val Acc: {:4f}".format(best_acc))# 训练完后用最好的一次当做模型最终的结果model.load_state_dict(best_model_weights)# 返回return model, val_acc_history, train_acc_history, valid_losses, train_losses, LRsdef get_data():"""获取数据"""# 获取测试集train = torchvision.datasets.CIFAR100(root="./mnt", train=True, download=True,transform=torchvision.transforms.Compose([torchvision.transforms.ToTensor(), # 转换成张量torchvision.transforms.Normalize((0.1307,), (0.3081,)) # 标准化]))train_loader = DataLoader(train, batch_size=batch_size) # 分割测试集# 获取测试集test = torchvision.datasets.CIFAR100(root="./mnt", train=False, download=True,transform=torchvision.transforms.Compose([torchvision.transforms.ToTensor(), # 转换成张量torchvision.transforms.Normalize((0.1307,), (0.3081,)) # 标准化]))test_loader = DataLoader(test, batch_size=batch_size) # 分割训练data_loader = {"train": train_loader, "valid": test_loader}# 返回分割好的训练集和测试集return data_loader# 超参数
filename = "checkpoint.pth" # 模型保存
feature_exact = True # 冻层
num_classes = 100 # 输出的类别数
batch_size = 1024 # 一次训练的样本数目
iteration_num = 10 # 迭代次数# 获取模型
resnet152 = initialize_model(model_name="resnet",num_classes=num_classes,feature_exact=feature_exact,use_pretrained=True
)# 是否使用GPU训练
use_cuda = torch.cuda.is_available()
device = torch.device("cuda" if use_cuda else "cpu")
if use_cuda: resnet152.cuda() # GPU 计算
print("是否使用 GPU 加速:", use_cuda)# 输出网络结构
print(summary(resnet152, (3, 32, 32)))# 训练参数
params_to_update = parameter_to_update(resnet152)# 优化器
optimizer = torch.optim.Adam(params_to_update, lr=0.01)
scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=10, gamma=0.1) # 学习率每10个epoch衰减到原来的1/10
criterion = torch.nn.NLLLoss()if __name__ == "__main__":data_loader = get_data()resnet152, val_acc_history, train_acc_history, valid_losses, train_losses, LRs = train_model(model=resnet152,dataloaders=data_loader,citerion=criterion,optimizer=optimizer,num_epochs=iteration_num,filename=filename)
输出结果:
是否使用 GPU 加速: True
----------------------------------------------------------------Layer (type) Output Shape Param #
================================================================Conv2d-1 [-1, 64, 16, 16] 9,408BatchNorm2d-2 [-1, 64, 16, 16] 128ReLU-3 [-1, 64, 16, 16] 0MaxPool2d-4 [-1, 64, 8, 8] 0Conv2d-5 [-1, 64, 8, 8] 4,096BatchNorm2d-6 [-1, 64, 8, 8] 128ReLU-7 [-1, 64, 8, 8] 0Conv2d-8 [-1, 64, 8, 8] 36,864BatchNorm2d-9 [-1, 64, 8, 8] 128ReLU-10 [-1, 64, 8, 8] 0Conv2d-11 [-1, 256, 8, 8] 16,384BatchNorm2d-12 [-1, 256, 8, 8] 512Conv2d-13 [-1, 256, 8, 8] 16,384BatchNorm2d-14 [-1, 256, 8, 8] 512ReLU-15 [-1, 256, 8, 8] 0Bottleneck-16 [-1, 256, 8, 8] 0Conv2d-17 [-1, 64, 8, 8] 16,384BatchNorm2d-18 [-1, 64, 8, 8] 128ReLU-19 [-1, 64, 8, 8] 0Conv2d-20 [-1, 64, 8, 8] 36,864BatchNorm2d-21 [-1, 64, 8, 8] 128ReLU-22 [-1, 64, 8, 8] 0Conv2d-23 [-1, 256, 8, 8] 16,384BatchNorm2d-24 [-1, 256, 8, 8] 512ReLU-25 [-1, 256, 8, 8] 0Bottleneck-26 [-1, 256, 8, 8] 0Conv2d-27 [-1, 64, 8, 8] 16,384BatchNorm2d-28 [-1, 64, 8, 8] 128ReLU-29 [-1, 64, 8, 8] 0Conv2d-30 [-1, 64, 8, 8] 36,864BatchNorm2d-31 [-1, 64, 8, 8] 128ReLU-32 [-1, 64, 8, 8] 0Conv2d-33 [-1, 256, 8, 8] 16,384BatchNorm2d-34 [-1, 256, 8, 8] 512ReLU-35 [-1, 256, 8, 8] 0Bottleneck-36 [-1, 256, 8, 8] 0Conv2d-37 [-1, 128, 8, 8] 32,768BatchNorm2d-38 [-1, 128, 8, 8] 256ReLU-39 [-1, 128, 8, 8] 0Conv2d-40 [-1, 128, 4, 4] 147,456BatchNorm2d-41 [-1, 128, 4, 4] 256ReLU-42 [-1, 128, 4, 4] 0Conv2d-43 [-1, 512, 4, 4] 65,536BatchNorm2d-44 [-1, 512, 4, 4] 1,024Conv2d-45 [-1, 512, 4, 4] 131,072BatchNorm2d-46 [-1, 512, 4, 4] 1,024ReLU-47 [-1, 512, 4, 4] 0Bottleneck-48 [-1, 512, 4, 4] 0Conv2d-49 [-1, 128, 4, 4] 65,536BatchNorm2d-50 [-1, 128, 4, 4] 256ReLU-51 [-1, 128, 4, 4] 0Conv2d-52 [-1, 128, 4, 4] 147,456BatchNorm2d-53 [-1, 128, 4, 4] 256ReLU-54 [-1, 128, 4, 4] 0Conv2d-55 [-1, 512, 4, 4] 65,536BatchNorm2d-56 [-1, 512, 4, 4] 1,024ReLU-57 [-1, 512, 4, 4] 0Bottleneck-58 [-1, 512, 4, 4] 0Conv2d-59 [-1, 128, 4, 4] 65,536BatchNorm2d-60 [-1, 128, 4, 4] 256ReLU-61 [-1, 128, 4, 4] 0Conv2d-62 [-1, 128, 4, 4] 147,456BatchNorm2d-63 [-1, 128, 4, 4] 256ReLU-64 [-1, 128, 4, 4] 0Conv2d-65 [-1, 512, 4, 4] 65,536BatchNorm2d-66 [-1, 512, 4, 4] 1,024ReLU-67 [-1, 512, 4, 4] 0Bottleneck-68 [-1, 512, 4, 4] 0Conv2d-69 [-1, 128, 4, 4] 65,536BatchNorm2d-70 [-1, 128, 4, 4] 256ReLU-71 [-1, 128, 4, 4] 0Conv2d-72 [-1, 128, 4, 4] 147,456BatchNorm2d-73 [-1, 128, 4, 4] 256ReLU-74 [-1, 128, 4, 4] 0Conv2d-75 [-1, 512, 4, 4] 65,536BatchNorm2d-76 [-1, 512, 4, 4] 1,024ReLU-77 [-1, 512, 4, 4] 0Bottleneck-78 [-1, 512, 4, 4] 0Conv2d-79 [-1, 128, 4, 4] 65,536BatchNorm2d-80 [-1, 128, 4, 4] 256ReLU-81 [-1, 128, 4, 4] 0Conv2d-82 [-1, 128, 4, 4] 147,456BatchNorm2d-83 [-1, 128, 4, 4] 256ReLU-84 [-1, 128, 4, 4] 0Conv2d-85 [-1, 512, 4, 4] 65,536BatchNorm2d-86 [-1, 512, 4, 4] 1,024ReLU-87 [-1, 512, 4, 4] 0Bottleneck-88 [-1, 512, 4, 4] 0Conv2d-89 [-1, 128, 4, 4] 65,536BatchNorm2d-90 [-1, 128, 4, 4] 256ReLU-91 [-1, 128, 4, 4] 0Conv2d-92 [-1, 128, 4, 4] 147,456BatchNorm2d-93 [-1, 128, 4, 4] 256ReLU-94 [-1, 128, 4, 4] 0Conv2d-95 [-1, 512, 4, 4] 65,536BatchNorm2d-96 [-1, 512, 4, 4] 1,024ReLU-97 [-1, 512, 4, 4] 0Bottleneck-98 [-1, 512, 4, 4] 0Conv2d-99 [-1, 128, 4, 4] 65,536BatchNorm2d-100 [-1, 128, 4, 4] 256ReLU-101 [-1, 128, 4, 4] 0Conv2d-102 [-1, 128, 4, 4] 147,456BatchNorm2d-103 [-1, 128, 4, 4] 256ReLU-104 [-1, 128, 4, 4] 0Conv2d-105 [-1, 512, 4, 4] 65,536BatchNorm2d-106 [-1, 512, 4, 4] 1,024ReLU-107 [-1, 512, 4, 4] 0Bottleneck-108 [-1, 512, 4, 4] 0Conv2d-109 [-1, 128, 4, 4] 65,536BatchNorm2d-110 [-1, 128, 4, 4] 256ReLU-111 [-1, 128, 4, 4] 0Conv2d-112 [-1, 128, 4, 4] 147,456BatchNorm2d-113 [-1, 128, 4, 4] 256ReLU-114 [-1, 128, 4, 4] 0Conv2d-115 [-1, 512, 4, 4] 65,536BatchNorm2d-116 [-1, 512, 4, 4] 1,024ReLU-117 [-1, 512, 4, 4] 0Bottleneck-118 [-1, 512, 4, 4] 0Conv2d-119 [-1, 256, 4, 4] 131,072BatchNorm2d-120 [-1, 256, 4, 4] 512ReLU-121 [-1, 256, 4, 4] 0Conv2d-122 [-1, 256, 2, 2] 589,824BatchNorm2d-123 [-1, 256, 2, 2] 512ReLU-124 [-1, 256, 2, 2] 0Conv2d-125 [-1, 1024, 2, 2] 262,144BatchNorm2d-126 [-1, 1024, 2, 2] 2,048Conv2d-127 [-1, 1024, 2, 2] 524,288BatchNorm2d-128 [-1, 1024, 2, 2] 2,048ReLU-129 [-1, 1024, 2, 2] 0Bottleneck-130 [-1, 1024, 2, 2] 0Conv2d-131 [-1, 256, 2, 2] 262,144BatchNorm2d-132 [-1, 256, 2, 2] 512ReLU-133 [-1, 256, 2, 2] 0Conv2d-134 [-1, 256, 2, 2] 589,824BatchNorm2d-135 [-1, 256, 2, 2] 512ReLU-136 [-1, 256, 2, 2] 0Conv2d-137 [-1, 1024, 2, 2] 262,144BatchNorm2d-138 [-1, 1024, 2, 2] 2,048ReLU-139 [-1, 1024, 2, 2] 0Bottleneck-140 [-1, 1024, 2, 2] 0Conv2d-141 [-1, 256, 2, 2] 262,144BatchNorm2d-142 [-1, 256, 2, 2] 512ReLU-143 [-1, 256, 2, 2] 0Conv2d-144 [-1, 256, 2, 2] 589,824BatchNorm2d-145 [-1, 256, 2, 2] 512ReLU-146 [-1, 256, 2, 2] 0Conv2d-147 [-1, 1024, 2, 2] 262,144BatchNorm2d-148 [-1, 1024, 2, 2] 2,048ReLU-149 [-1, 1024, 2, 2] 0Bottleneck-150 [-1, 1024, 2, 2] 0Conv2d-151 [-1, 256, 2, 2] 262,144BatchNorm2d-152 [-1, 256, 2, 2] 512ReLU-153 [-1, 256, 2, 2] 0Conv2d-154 [-1, 256, 2, 2] 589,824BatchNorm2d-155 [-1, 256, 2, 2] 512ReLU-156 [-1, 256, 2, 2] 0Conv2d-157 [-1, 1024, 2, 2] 262,144BatchNorm2d-158 [-1, 1024, 2, 2] 2,048ReLU-159 [-1, 1024, 2, 2] 0Bottleneck-160 [-1, 1024, 2, 2] 0Conv2d-161 [-1, 256, 2, 2] 262,144BatchNorm2d-162 [-1, 256, 2, 2] 512ReLU-163 [-1, 256, 2, 2] 0Conv2d-164 [-1, 256, 2, 2] 589,824BatchNorm2d-165 [-1, 256, 2, 2] 512ReLU-166 [-1, 256, 2, 2] 0Conv2d-167 [-1, 1024, 2, 2] 262,144BatchNorm2d-168 [-1, 1024, 2, 2] 2,048ReLU-169 [-1, 1024, 2, 2] 0Bottleneck-170 [-1, 1024, 2, 2] 0Conv2d-171 [-1, 256, 2, 2] 262,144BatchNorm2d-172 [-1, 256, 2, 2] 512ReLU-173 [-1, 256, 2, 2] 0Conv2d-174 [-1, 256, 2, 2] 589,824BatchNorm2d-175 [-1, 256, 2, 2] 512ReLU-176 [-1, 256, 2, 2] 0Conv2d-177 [-1, 1024, 2, 2] 262,144BatchNorm2d-178 [-1, 1024, 2, 2] 2,048ReLU-179 [-1, 1024, 2, 2] 0Bottleneck-180 [-1, 1024, 2, 2] 0Conv2d-181 [-1, 256, 2, 2] 262,144BatchNorm2d-182 [-1, 256, 2, 2] 512ReLU-183 [-1, 256, 2, 2] 0Conv2d-184 [-1, 256, 2, 2] 589,824BatchNorm2d-185 [-1, 256, 2, 2] 512ReLU-186 [-1, 256, 2, 2] 0Conv2d-187 [-1, 1024, 2, 2] 262,144BatchNorm2d-188 [-1, 1024, 2, 2] 2,048ReLU-189 [-1, 1024, 2, 2] 0Bottleneck-190 [-1, 1024, 2, 2] 0Conv2d-191 [-1, 256, 2, 2] 262,144BatchNorm2d-192 [-1, 256, 2, 2] 512ReLU-193 [-1, 256, 2, 2] 0Conv2d-194 [-1, 256, 2, 2] 589,824BatchNorm2d-195 [-1, 256, 2, 2] 512ReLU-196 [-1, 256, 2, 2] 0Conv2d-197 [-1, 1024, 2, 2] 262,144BatchNorm2d-198 [-1, 1024, 2, 2] 2,048ReLU-199 [-1, 1024, 2, 2] 0Bottleneck-200 [-1, 1024, 2, 2] 0Conv2d-201 [-1, 256, 2, 2] 262,144BatchNorm2d-202 [-1, 256, 2, 2] 512ReLU-203 [-1, 256, 2, 2] 0Conv2d-204 [-1, 256, 2, 2] 589,824BatchNorm2d-205 [-1, 256, 2, 2] 512ReLU-206 [-1, 256, 2, 2] 0Conv2d-207 [-1, 1024, 2, 2] 262,144BatchNorm2d-208 [-1, 1024, 2, 2] 2,048ReLU-209 [-1, 1024, 2, 2] 0Bottleneck-210 [-1, 1024, 2, 2] 0Conv2d-211 [-1, 256, 2, 2] 262,144BatchNorm2d-212 [-1, 256, 2, 2] 512ReLU-213 [-1, 256, 2, 2] 0Conv2d-214 [-1, 256, 2, 2] 589,824BatchNorm2d-215 [-1, 256, 2, 2] 512ReLU-216 [-1, 256, 2, 2] 0Conv2d-217 [-1, 1024, 2, 2] 262,144BatchNorm2d-218 [-1, 1024, 2, 2] 2,048ReLU-219 [-1, 1024, 2, 2] 0Bottleneck-220 [-1, 1024, 2, 2] 0Conv2d-221 [-1, 256, 2, 2] 262,144BatchNorm2d-222 [-1, 256, 2, 2] 512ReLU-223 [-1, 256, 2, 2] 0Conv2d-224 [-1, 256, 2, 2] 589,824BatchNorm2d-225 [-1, 256, 2, 2] 512ReLU-226 [-1, 256, 2, 2] 0Conv2d-227 [-1, 1024, 2, 2] 262,144BatchNorm2d-228 [-1, 1024, 2, 2] 2,048ReLU-229 [-1, 1024, 2, 2] 0Bottleneck-230 [-1, 1024, 2, 2] 0Conv2d-231 [-1, 256, 2, 2] 262,144BatchNorm2d-232 [-1, 256, 2, 2] 512ReLU-233 [-1, 256, 2, 2] 0Conv2d-234 [-1, 256, 2, 2] 589,824BatchNorm2d-235 [-1, 256, 2, 2] 512ReLU-236 [-1, 256, 2, 2] 0Conv2d-237 [-1, 1024, 2, 2] 262,144BatchNorm2d-238 [-1, 1024, 2, 2] 2,048ReLU-239 [-1, 1024, 2, 2] 0Bottleneck-240 [-1, 1024, 2, 2] 0Conv2d-241 [-1, 256, 2, 2] 262,144BatchNorm2d-242 [-1, 256, 2, 2] 512ReLU-243 [-1, 256, 2, 2] 0Conv2d-244 [-1, 256, 2, 2] 589,824BatchNorm2d-245 [-1, 256, 2, 2] 512ReLU-246 [-1, 256, 2, 2] 0Conv2d-247 [-1, 1024, 2, 2] 262,144BatchNorm2d-248 [-1, 1024, 2, 2] 2,048ReLU-249 [-1, 1024, 2, 2] 0Bottleneck-250 [-1, 1024, 2, 2] 0Conv2d-251 [-1, 256, 2, 2] 262,144BatchNorm2d-252 [-1, 256, 2, 2] 512ReLU-253 [-1, 256, 2, 2] 0Conv2d-254 [-1, 256, 2, 2] 589,824BatchNorm2d-255 [-1, 256, 2, 2] 512ReLU-256 [-1, 256, 2, 2] 0Conv2d-257 [-1, 1024, 2, 2] 262,144BatchNorm2d-258 [-1, 1024, 2, 2] 2,048ReLU-259 [-1, 1024, 2, 2] 0Bottleneck-260 [-1, 1024, 2, 2] 0Conv2d-261 [-1, 256, 2, 2] 262,144BatchNorm2d-262 [-1, 256, 2, 2] 512ReLU-263 [-1, 256, 2, 2] 0Conv2d-264 [-1, 256, 2, 2] 589,824BatchNorm2d-265 [-1, 256, 2, 2] 512ReLU-266 [-1, 256, 2, 2] 0Conv2d-267 [-1, 1024, 2, 2] 262,144BatchNorm2d-268 [-1, 1024, 2, 2] 2,048ReLU-269 [-1, 1024, 2, 2] 0Bottleneck-270 [-1, 1024, 2, 2] 0Conv2d-271 [-1, 256, 2, 2] 262,144BatchNorm2d-272 [-1, 256, 2, 2] 512ReLU-273 [-1, 256, 2, 2] 0Conv2d-274 [-1, 256, 2, 2] 589,824BatchNorm2d-275 [-1, 256, 2, 2] 512ReLU-276 [-1, 256, 2, 2] 0Conv2d-277 [-1, 1024, 2, 2] 262,144BatchNorm2d-278 [-1, 1024, 2, 2] 2,048ReLU-279 [-1, 1024, 2, 2] 0Bottleneck-280 [-1, 1024, 2, 2] 0Conv2d-281 [-1, 256, 2, 2] 262,144BatchNorm2d-282 [-1, 256, 2, 2] 512ReLU-283 [-1, 256, 2, 2] 0Conv2d-284 [-1, 256, 2, 2] 589,824BatchNorm2d-285 [-1, 256, 2, 2] 512ReLU-286 [-1, 256, 2, 2] 0Conv2d-287 [-1, 1024, 2, 2] 262,144BatchNorm2d-288 [-1, 1024, 2, 2] 2,048ReLU-289 [-1, 1024, 2, 2] 0Bottleneck-290 [-1, 1024, 2, 2] 0Conv2d-291 [-1, 256, 2, 2] 262,144BatchNorm2d-292 [-1, 256, 2, 2] 512ReLU-293 [-1, 256, 2, 2] 0Conv2d-294 [-1, 256, 2, 2] 589,824BatchNorm2d-295 [-1, 256, 2, 2] 512ReLU-296 [-1, 256, 2, 2] 0Conv2d-297 [-1, 1024, 2, 2] 262,144BatchNorm2d-298 [-1, 1024, 2, 2] 2,048ReLU-299 [-1, 1024, 2, 2] 0Bottleneck-300 [-1, 1024, 2, 2] 0Conv2d-301 [-1, 256, 2, 2] 262,144BatchNorm2d-302 [-1, 256, 2, 2] 512ReLU-303 [-1, 256, 2, 2] 0Conv2d-304 [-1, 256, 2, 2] 589,824BatchNorm2d-305 [-1, 256, 2, 2] 512ReLU-306 [-1, 256, 2, 2] 0Conv2d-307 [-1, 1024, 2, 2] 262,144BatchNorm2d-308 [-1, 1024, 2, 2] 2,048ReLU-309 [-1, 1024, 2, 2] 0Bottleneck-310 [-1, 1024, 2, 2] 0Conv2d-311 [-1, 256, 2, 2] 262,144BatchNorm2d-312 [-1, 256, 2, 2] 512ReLU-313 [-1, 256, 2, 2] 0Conv2d-314 [-1, 256, 2, 2] 589,824BatchNorm2d-315 [-1, 256, 2, 2] 512ReLU-316 [-1, 256, 2, 2] 0Conv2d-317 [-1, 1024, 2, 2] 262,144BatchNorm2d-318 [-1, 1024, 2, 2] 2,048ReLU-319 [-1, 1024, 2, 2] 0Bottleneck-320 [-1, 1024, 2, 2] 0Conv2d-321 [-1, 256, 2, 2] 262,144BatchNorm2d-322 [-1, 256, 2, 2] 512ReLU-323 [-1, 256, 2, 2] 0Conv2d-324 [-1, 256, 2, 2] 589,824BatchNorm2d-325 [-1, 256, 2, 2] 512ReLU-326 [-1, 256, 2, 2] 0Conv2d-327 [-1, 1024, 2, 2] 262,144BatchNorm2d-328 [-1, 1024, 2, 2] 2,048ReLU-329 [-1, 1024, 2, 2] 0Bottleneck-330 [-1, 1024, 2, 2] 0Conv2d-331 [-1, 256, 2, 2] 262,144BatchNorm2d-332 [-1, 256, 2, 2] 512ReLU-333 [-1, 256, 2, 2] 0Conv2d-334 [-1, 256, 2, 2] 589,824BatchNorm2d-335 [-1, 256, 2, 2] 512ReLU-336 [-1, 256, 2, 2] 0Conv2d-337 [-1, 1024, 2, 2] 262,144BatchNorm2d-338 [-1, 1024, 2, 2] 2,048ReLU-339 [-1, 1024, 2, 2] 0Bottleneck-340 [-1, 1024, 2, 2] 0Conv2d-341 [-1, 256, 2, 2] 262,144BatchNorm2d-342 [-1, 256, 2, 2] 512ReLU-343 [-1, 256, 2, 2] 0Conv2d-344 [-1, 256, 2, 2] 589,824BatchNorm2d-345 [-1, 256, 2, 2] 512ReLU-346 [-1, 256, 2, 2] 0Conv2d-347 [-1, 1024, 2, 2] 262,144BatchNorm2d-348 [-1, 1024, 2, 2] 2,048ReLU-349 [-1, 1024, 2, 2] 0Bottleneck-350 [-1, 1024, 2, 2] 0Conv2d-351 [-1, 256, 2, 2] 262,144BatchNorm2d-352 [-1, 256, 2, 2] 512ReLU-353 [-1, 256, 2, 2] 0Conv2d-354 [-1, 256, 2, 2] 589,824BatchNorm2d-355 [-1, 256, 2, 2] 512ReLU-356 [-1, 256, 2, 2] 0Conv2d-357 [-1, 1024, 2, 2] 262,144BatchNorm2d-358 [-1, 1024, 2, 2] 2,048ReLU-359 [-1, 1024, 2, 2] 0Bottleneck-360 [-1, 1024, 2, 2] 0Conv2d-361 [-1, 256, 2, 2] 262,144BatchNorm2d-362 [-1, 256, 2, 2] 512ReLU-363 [-1, 256, 2, 2] 0Conv2d-364 [-1, 256, 2, 2] 589,824BatchNorm2d-365 [-1, 256, 2, 2] 512ReLU-366 [-1, 256, 2, 2] 0Conv2d-367 [-1, 1024, 2, 2] 262,144BatchNorm2d-368 [-1, 1024, 2, 2] 2,048ReLU-369 [-1, 1024, 2, 2] 0Bottleneck-370 [-1, 1024, 2, 2] 0Conv2d-371 [-1, 256, 2, 2] 262,144BatchNorm2d-372 [-1, 256, 2, 2] 512ReLU-373 [-1, 256, 2, 2] 0Conv2d-374 [-1, 256, 2, 2] 589,824BatchNorm2d-375 [-1, 256, 2, 2] 512ReLU-376 [-1, 256, 2, 2] 0Conv2d-377 [-1, 1024, 2, 2] 262,144BatchNorm2d-378 [-1, 1024, 2, 2] 2,048ReLU-379 [-1, 1024, 2, 2] 0Bottleneck-380 [-1, 1024, 2, 2] 0Conv2d-381 [-1, 256, 2, 2] 262,144BatchNorm2d-382 [-1, 256, 2, 2] 512ReLU-383 [-1, 256, 2, 2] 0Conv2d-384 [-1, 256, 2, 2] 589,824BatchNorm2d-385 [-1, 256, 2, 2] 512ReLU-386 [-1, 256, 2, 2] 0Conv2d-387 [-1, 1024, 2, 2] 262,144BatchNorm2d-388 [-1, 1024, 2, 2] 2,048ReLU-389 [-1, 1024, 2, 2] 0Bottleneck-390 [-1, 1024, 2, 2] 0Conv2d-391 [-1, 256, 2, 2] 262,144BatchNorm2d-392 [-1, 256, 2, 2] 512ReLU-393 [-1, 256, 2, 2] 0Conv2d-394 [-1, 256, 2, 2] 589,824BatchNorm2d-395 [-1, 256, 2, 2] 512ReLU-396 [-1, 256, 2, 2] 0Conv2d-397 [-1, 1024, 2, 2] 262,144BatchNorm2d-398 [-1, 1024, 2, 2] 2,048ReLU-399 [-1, 1024, 2, 2] 0Bottleneck-400 [-1, 1024, 2, 2] 0Conv2d-401 [-1, 256, 2, 2] 262,144BatchNorm2d-402 [-1, 256, 2, 2] 512ReLU-403 [-1, 256, 2, 2] 0Conv2d-404 [-1, 256, 2, 2] 589,824BatchNorm2d-405 [-1, 256, 2, 2] 512ReLU-406 [-1, 256, 2, 2] 0Conv2d-407 [-1, 1024, 2, 2] 262,144BatchNorm2d-408 [-1, 1024, 2, 2] 2,048ReLU-409 [-1, 1024, 2, 2] 0Bottleneck-410 [-1, 1024, 2, 2] 0Conv2d-411 [-1, 256, 2, 2] 262,144BatchNorm2d-412 [-1, 256, 2, 2] 512ReLU-413 [-1, 256, 2, 2] 0Conv2d-414 [-1, 256, 2, 2] 589,824BatchNorm2d-415 [-1, 256, 2, 2] 512ReLU-416 [-1, 256, 2, 2] 0Conv2d-417 [-1, 1024, 2, 2] 262,144BatchNorm2d-418 [-1, 1024, 2, 2] 2,048ReLU-419 [-1, 1024, 2, 2] 0Bottleneck-420 [-1, 1024, 2, 2] 0Conv2d-421 [-1, 256, 2, 2] 262,144BatchNorm2d-422 [-1, 256, 2, 2] 512ReLU-423 [-1, 256, 2, 2] 0Conv2d-424 [-1, 256, 2, 2] 589,824BatchNorm2d-425 [-1, 256, 2, 2] 512ReLU-426 [-1, 256, 2, 2] 0Conv2d-427 [-1, 1024, 2, 2] 262,144BatchNorm2d-428 [-1, 1024, 2, 2] 2,048ReLU-429 [-1, 1024, 2, 2] 0Bottleneck-430 [-1, 1024, 2, 2] 0Conv2d-431 [-1, 256, 2, 2] 262,144BatchNorm2d-432 [-1, 256, 2, 2] 512ReLU-433 [-1, 256, 2, 2] 0Conv2d-434 [-1, 256, 2, 2] 589,824BatchNorm2d-435 [-1, 256, 2, 2] 512ReLU-436 [-1, 256, 2, 2] 0Conv2d-437 [-1, 1024, 2, 2] 262,144BatchNorm2d-438 [-1, 1024, 2, 2] 2,048ReLU-439 [-1, 1024, 2, 2] 0Bottleneck-440 [-1, 1024, 2, 2] 0Conv2d-441 [-1, 256, 2, 2] 262,144BatchNorm2d-442 [-1, 256, 2, 2] 512ReLU-443 [-1, 256, 2, 2] 0Conv2d-444 [-1, 256, 2, 2] 589,824BatchNorm2d-445 [-1, 256, 2, 2] 512ReLU-446 [-1, 256, 2, 2] 0Conv2d-447 [-1, 1024, 2, 2] 262,144BatchNorm2d-448 [-1, 1024, 2, 2] 2,048ReLU-449 [-1, 1024, 2, 2] 0Bottleneck-450 [-1, 1024, 2, 2] 0Conv2d-451 [-1, 256, 2, 2] 262,144BatchNorm2d-452 [-1, 256, 2, 2] 512ReLU-453 [-1, 256, 2, 2] 0Conv2d-454 [-1, 256, 2, 2] 589,824BatchNorm2d-455 [-1, 256, 2, 2] 512ReLU-456 [-1, 256, 2, 2] 0Conv2d-457 [-1, 1024, 2, 2] 262,144BatchNorm2d-458 [-1, 1024, 2, 2] 2,048ReLU-459 [-1, 1024, 2, 2] 0Bottleneck-460 [-1, 1024, 2, 2] 0Conv2d-461 [-1, 256, 2, 2] 262,144BatchNorm2d-462 [-1, 256, 2, 2] 512ReLU-463 [-1, 256, 2, 2] 0Conv2d-464 [-1, 256, 2, 2] 589,824BatchNorm2d-465 [-1, 256, 2, 2] 512ReLU-466 [-1, 256, 2, 2] 0Conv2d-467 [-1, 1024, 2, 2] 262,144BatchNorm2d-468 [-1, 1024, 2, 2] 2,048ReLU-469 [-1, 1024, 2, 2] 0Bottleneck-470 [-1, 1024, 2, 2] 0Conv2d-471 [-1, 256, 2, 2] 262,144BatchNorm2d-472 [-1, 256, 2, 2] 512ReLU-473 [-1, 256, 2, 2] 0Conv2d-474 [-1, 256, 2, 2] 589,824BatchNorm2d-475 [-1, 256, 2, 2] 512ReLU-476 [-1, 256, 2, 2] 0Conv2d-477 [-1, 1024, 2, 2] 262,144BatchNorm2d-478 [-1, 1024, 2, 2] 2,048ReLU-479 [-1, 1024, 2, 2] 0Bottleneck-480 [-1, 1024, 2, 2] 0Conv2d-481 [-1, 512, 2, 2] 524,288BatchNorm2d-482 [-1, 512, 2, 2] 1,024ReLU-483 [-1, 512, 2, 2] 0Conv2d-484 [-1, 512, 1, 1] 2,359,296BatchNorm2d-485 [-1, 512, 1, 1] 1,024ReLU-486 [-1, 512, 1, 1] 0Conv2d-487 [-1, 2048, 1, 1] 1,048,576BatchNorm2d-488 [-1, 2048, 1, 1] 4,096Conv2d-489 [-1, 2048, 1, 1] 2,097,152BatchNorm2d-490 [-1, 2048, 1, 1] 4,096ReLU-491 [-1, 2048, 1, 1] 0Bottleneck-492 [-1, 2048, 1, 1] 0Conv2d-493 [-1, 512, 1, 1] 1,048,576BatchNorm2d-494 [-1, 512, 1, 1] 1,024ReLU-495 [-1, 512, 1, 1] 0Conv2d-496 [-1, 512, 1, 1] 2,359,296BatchNorm2d-497 [-1, 512, 1, 1] 1,024ReLU-498 [-1, 512, 1, 1] 0Conv2d-499 [-1, 2048, 1, 1] 1,048,576BatchNorm2d-500 [-1, 2048, 1, 1] 4,096ReLU-501 [-1, 2048, 1, 1] 0Bottleneck-502 [-1, 2048, 1, 1] 0Conv2d-503 [-1, 512, 1, 1] 1,048,576BatchNorm2d-504 [-1, 512, 1, 1] 1,024ReLU-505 [-1, 512, 1, 1] 0Conv2d-506 [-1, 512, 1, 1] 2,359,296BatchNorm2d-507 [-1, 512, 1, 1] 1,024ReLU-508 [-1, 512, 1, 1] 0Conv2d-509 [-1, 2048, 1, 1] 1,048,576BatchNorm2d-510 [-1, 2048, 1, 1] 4,096ReLU-511 [-1, 2048, 1, 1] 0Bottleneck-512 [-1, 2048, 1, 1] 0
AdaptiveAvgPool2d-513 [-1, 2048, 1, 1] 0Linear-514 [-1, 100] 204,900LogSoftmax-515 [-1, 100] 0
================================================================
Total params: 58,348,708
Trainable params: 204,900
Non-trainable params: 58,143,808
----------------------------------------------------------------
Input size (MB): 0.01
Forward/backward pass size (MB): 12.40
Params size (MB): 222.58
Estimated Total Size (MB): 234.99
----------------------------------------------------------------
None
Params to learnfc.0.weightfc.0.bias
Files already downloaded and verified
Files already downloaded and verified
Epoch 0/9
----------
Time elapsed 0m 21s
train Loss: 7.5111 Acc: 0.1484
Time elapsed 0m 26s
valid Loss: 3.7821 Acc: 0.2493
/usr/local/lib/python3.7/dist-packages/torch/optim/lr_scheduler.py:154: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
Optimizer learning rate: 0.0100000Epoch 1/9
----------
Time elapsed 0m 47s
train Loss: 2.9405 Acc: 0.3109
Time elapsed 0m 52s
valid Loss: 3.2014 Acc: 0.2739
Optimizer learning rate: 0.0100000Epoch 2/9
----------
Time elapsed 1m 12s
train Loss: 2.5866 Acc: 0.3622
Time elapsed 1m 17s
valid Loss: 3.2239 Acc: 0.2787
Optimizer learning rate: 0.0100000Epoch 3/9
----------
Time elapsed 1m 38s
train Loss: 2.4077 Acc: 0.3969
Time elapsed 1m 43s
valid Loss: 3.2608 Acc: 0.2811
Optimizer learning rate: 0.0100000Epoch 4/9
----------
Time elapsed 2m 4s
train Loss: 2.2742 Acc: 0.4263
Time elapsed 2m 9s
valid Loss: 3.4260 Acc: 0.2689
Optimizer learning rate: 0.0100000Epoch 5/9
----------
Time elapsed 2m 29s
train Loss: 2.1942 Acc: 0.4434
Time elapsed 2m 34s
valid Loss: 3.4697 Acc: 0.2760
Optimizer learning rate: 0.0100000Epoch 6/9
----------
Time elapsed 2m 54s
train Loss: 2.1369 Acc: 0.4583
Time elapsed 2m 59s
valid Loss: 3.5391 Acc: 0.2744
Optimizer learning rate: 0.0100000Epoch 7/9
----------
Time elapsed 3m 20s
train Loss: 2.0382 Acc: 0.4771
Time elapsed 3m 24s
valid Loss: 3.5992 Acc: 0.2721
Optimizer learning rate: 0.0100000Epoch 8/9
----------
Time elapsed 3m 45s
train Loss: 1.9776 Acc: 0.4939
Time elapsed 3m 50s
valid Loss: 3.7533 Acc: 0.2685
Optimizer learning rate: 0.0100000Epoch 9/9
----------
Time elapsed 4m 11s
train Loss: 1.9309 Acc: 0.5035
Time elapsed 4m 16s
valid Loss: 3.9663 Acc: 0.2558
Optimizer learning rate: 0.0100000Training complete in 4m 16s
Best val Acc: 0.281100
PyTorch 迁移学习 (Transfer Learning) 代码详解相关推荐
- 迁移学习(Transfer learning)、重用预训练图层、预训练模型库
迁移学习(Transfer learning).重用预训练图层.预训练模型库 目录 迁移学习(Transfer learning).重用预训练图层.预训练模型库 迁移学
- 深度学习不得不会的迁移学习Transfer Learning
http://blog.itpub.net/29829936/viewspace-2641919/ 2019-04-18 10:04:53 目录 一.概述 二.什么是迁移学习? 2.1 模型的训练与预 ...
- 迁移学习(transfer learning)与finetune的关系？【finetune只是transfer learning的一种手段】
目录 1.迁移学习简介 2.为什么要迁移学习? 3.迁移学习的几种方式 1)Transfer Learning: 2)Extract Feature Vector: 3)Fine-tune: 4.三种 ...
- 迁移学习(Transfer Learning)的背景、历史
迁移学习(Transfer Learning)的背景.历史 迁移学习的背景.历史及学习 1.迁移学习提出背景 在机器学习.深度学习和数据挖掘的大多数任务中,我们都会假设training和inferen ...
- 对比学习：MoCo代码详解
MoCo算法代码详解 本文代码来源: 1.导入包 2.参数设置 3.数据预处理 4. 模型 4.1moment update key encoder 4.2进队出队 4.3shuffle 4.4损失计 ...
- 【CV】Pytorch一小时入门教程-代码详解
目录 一.关键部分代码分解 1.定义网络 2.损失函数(代价函数) 3.更新权值 二.训练完整的分类器 1.数据处理 2. 训练模型(代码详解) CPU训练 GPU训练 CPU版本与GPU版本代码区别 ...
- 迁移学习 Transfer Learning—通俗易懂地介绍（常见网络模型pytorch实现）
前言 你会发现聪明人都喜欢"偷懒", 因为这样的偷懒能帮我们节省大量的时间, 提高效率. 还有一种偷懒是 "站在巨人的肩膀上". 不仅能看得更远, 还能看到更多 ...
- 迁移学习(Transfer Learning)概述及代码实现(full version)
基于PaddlePaddle的李宏毅机器学习--迁移学习 大噶好,我是黄波波.希望能和大家共进步,错误之处恳请指出! 百度AI Studio个人主页, 我在AI Studio上获得白银等级,点亮2个徽 ...
- 迁移学习(Transfer Learning)概述及代码实现
基于PaddlePaddle的李宏毅机器学习--迁移学习 大噶好,我是黄波波,希望能和大家共进步,错误之处恳请指出! 百度AI Studio个人主页, 我在AI Studio上获得白银等级,点亮2个徽 ...
最新文章
- 点击TableView任一行跳转详情页面会跳转两次的解决办法
- 如何写论文？看下这份《科研论文撰写策略》
- 4.4学习笔记-REGEXP1(正则表达式)
- lucence学习系列之一 基本概念
- spark 算子使用类变量_自己工作中超全spark性能优化总结
- 【Mybatis】 mapper XML 文件中使用 collection实现一对多关联
- 实验四Java_《Java实验四》
- 生产批量检查web url并通过手机邮件报警案例
- sqlserver 2008express版本启用混合登陆和sa
- bootstrap ie兼容
- SIFT算法学习总结
- ShellExecuteEx and SHELLEXECUTEINFO and ShellExecute
- Eplan PLC连接点-两两相连接方法
- 使用对话框模板创建一个InputBox()在C + +
- android监听系统来电并弹出提示窗口,Android 监听来去电 弹出悬浮窗提示
- SQL语句的五大类：DQL、DML、DDL、DTL、DCL
- Android使用自带的文件管理器选择文件并读取内容
- QT访问http服务器，并加载服务器图片
- PLC的面向对象编程
- 苹果X屏黑了-苹果x屏幕黑屏也无法开机怎么回事？
热门文章
- 安卓手机应用软件UI界面设计展示PR样机模板MOGRT
- python求定积分程序_在python中用sympy求定积分失败
- 微服务ServiceComb Meetup，携手Apache和华为大咖，分享微服务创新实践(上海，9月20日)...
- Kaggle入门准备与上网指南
- mysql 存储模式_FriendFeed 如何使用 MySQL 来存储无模式的数据
- JAVA面向对象----封装
- DCGAN---生成动漫头像
- 《C#零基础入门之百识百例》（六）类型转换 -- 转换用户输入
- 在 Windows 上安装 SVN Subversion 服务端
- 计数器代码php,计数器代码_php计数器的简单代码举例