Pytorch之迁移学习

迁移学习解决的问题:
1.实际任务中，很难得到一个数据量足够大的数据集，而且从0开始训练网络，消耗大量资源。
2.现有网络只针对某一特定问题，迁移学习将所学的知识，迁移到新场景，可以很好的解决其他相似问题。
迁移学习主要场景：
1.网络微调：使用预训练的网络(如在imagenet 1000上训练而来的网络)来初始化自己的网络，而不是随机初始化。将网络输出由1000改为2，以解决二分类问题。
2.将Convnet看成固定的特征提取器:首先固定ConvNet除了最后的全连接层外的其他所有层。最后的全连接层被替换成一个新的随机 初始化的层，只有这个新的层会被训练[只有这层参数会在反向传播时更新]

下面是利用PyTorch进行迁移学习步骤，要解决的问题是训练一个模型来对蚂蚁和蜜蜂进行分类。
1.导入相关包

import torch
import torch.nn as nn
import torch.optim as optim
from torch.optim import lr_scheduler
import numpy as np
import torchvision
from torchvision import datasets, models, transforms
import matplotlib.pyplot as plt
import time
import os
import copy
plt.ion()   # interactive mode

2.加载数据集

＃训练集数据扩充和归一化
＃在验证集上仅需要归一化
data_transforms = {
    'train': transforms.Compose([
        transforms.RandomResizedCrop(224), #随机裁剪一个area然后再resize
        transforms.RandomHorizontalFlip(), #随机水平翻转
        transforms.ToTensor(),
        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
    ]),
    'val': transforms.Compose([
        transforms.Resize(256),
        transforms.CenterCrop(224),
        transforms.ToTensor(),
        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
    ]),
}

data_dir = 'data/hymenoptera_data'
image_datasets = {x: datasets.ImageFolder(os.path.join(data_dir, x),
                                          data_transforms[x])
                  for x in ['train', 'val']}# 存在train和val两个文件夹
dataloaders = {x: torch.utils.data.DataLoader(image_datasets[x], batch_size=4,
                                             shuffle=True, num_workers=4)
              for x in ['train', 'val']}
dataset_sizes = {x: len(image_datasets[x]) for x in ['train', 'val']}
class_names = image_datasets['train'].classes

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

3.可视化部分图像数据

def imshow(inp, title=None):
    """Imshow for Tensor."""
    inp = inp.numpy().transpose((1, 2, 0))
    mean = np.array([0.485, 0.456, 0.406])
    std = np.array([0.229, 0.224, 0.225])
    inp = std * inp + mean
    inp = np.clip(inp, 0, 1)
    plt.imshow(inp)
    if title is not None:
        plt.title(title)
    plt.pause(0.001)  # pause a bit so that plots are updated


# 获取一批训练数据
inputs, classes = next(iter(dataloaders['train']))

# 批量制作网格
out = torchvision.utils.make_grid(inputs)

imshow(out, title=[class_names[x] for x in classes])

4.训练模型

def train_model(model, criterion, optimizer, scheduler, num_epochs=25):
    since = time.time()

    best_model_wts = copy.deepcopy(model.state_dict())
    best_acc = 0.0

    for epoch in range(num_epochs):
        print('Epoch {}/{}'.format(epoch, num_epochs - 1))
        print('-' * 10)

        # 每个epoch都有一个训练和验证阶段
        for phase in ['train', 'val']:
            if phase == 'train':
                scheduler.step()
                model.train()  # Set model to training mode
            else:
                model.eval()   # Set model to evaluate mode

            running_loss = 0.0
            running_corrects = 0

            # 迭代数据.
            for inputs, labels in dataloaders[phase]:
                inputs = inputs.to(device)
                labels = labels.to(device)

                # 零参数梯度
                optimizer.zero_grad()

                # 前向
                # track history if only in train
                with torch.set_grad_enabled(phase == 'train'):
                    outputs = model(inputs)
                    _, preds = torch.max(outputs, 1)
                    loss = criterion(outputs, labels)

                    # 后向+仅在训练阶段进行优化
                    if phase == 'train':
                        loss.backward()
                        optimizer.step()

                # 统计
                running_loss += loss.item() * inputs.size(0)
                running_corrects += torch.sum(preds == labels.data)

            epoch_loss = running_loss / dataset_sizes[phase]
            epoch_acc = running_corrects.double() / dataset_sizes[phase]

            print('{} Loss: {:.4f} Acc: {:.4f}'.format(
                phase, epoch_loss, epoch_acc))

            # 深度复制mo
            if phase == 'val' and epoch_acc > best_acc:
                best_acc = epoch_acc
                best_model_wts = copy.deepcopy(model.state_dict())

        print()

    time_elapsed = time.time() - since
    print('Training complete in {:.0f}m {:.0f}s'.format(
        time_elapsed // 60, time_elapsed % 60))
    print('Best val Acc: {:4f}'.format(best_acc))

    # 加载最佳模型权重
    model.load_state_dict(best_model_wts)
    return model

5.可视化模型的预测结果

#一个通用的展示少量预测图片的函数
def visualize_model(model, num_images=6):
    was_training = model.training
    model.eval()
    images_so_far = 0
    fig = plt.figure()

    with torch.no_grad():
        for i, (inputs, labels) in enumerate(dataloaders['val']):
            inputs = inputs.to(device)
            labels = labels.to(device)

            outputs = model(inputs)
            _, preds = torch.max(outputs, 1)

            for j in range(inputs.size()[0]):
                images_so_far += 1
                ax = plt.subplot(num_images//2, 2, images_so_far)
                ax.axis('off')
                ax.set_title('predicted: {}'.format(class_names[preds[j]]))
                imshow(inputs.cpu().data[j])

                if images_so_far == num_images:
                    model.train(mode=was_training)
                    return
        model.train(mode=was_training)

场景1：微调

# 加载预训练模型
model_ft = models.resnet18(pretrained=True)
# 获取全连接层输入数据
num_ftrs = model_ft.fc.in_features
# 修改最后一层全连接层，输出为2
model_ft.fc = nn.Linear(num_ftrs, 2)
# 将模型部署到GPU或CPU
model_ft = model_ft.to(device)

criterion = nn.CrossEntropyLoss()

# 观察所有参数都正在优化
optimizer_ft = optim.SGD(model_ft.parameters(), lr=0.001, momentum=0.9)

# 每7个epochs衰减LR通过设置gamma=0.1
exp_lr_scheduler = lr_scheduler.StepLR(optimizer_ft, step_size=7, gamma=0.1)
# 模型评估
model_ft = train_model(model_ft, criterion, optimizer_ft, exp_lr_scheduler, num_epochs=25)
# 输出
Epoch 0/24
----------
train Loss: 0.5503 Acc: 0.7213
val Loss: 0.2309 Acc: 0.9020

Epoch 1/24
----------
train Loss: 0.4198 Acc: 0.8115
val Loss: 0.2272 Acc: 0.9216
...
Epoch 24/24
----------
train Loss: 0.2832 Acc: 0.8730
val Loss: 0.2516 Acc: 0.9216

Training complete in 48m 20s
Best val Acc: 0.921569

场景2：ConvNet作为固定特征提取器
冻结除最后一层之外的所有网络。通过设置requires_grad == Falsebackward()来冻结参数，这样在反向传播backward()的时候他们的梯度就不会被计算。

model_conv = torchvision.models.resnet18(pretrained=True)
# 遍历冻结卷积层参数
for param in model_conv.parameters():
    param.requires_grad = False

# Parameters of newly constructed modules have requires_grad=True by default
num_ftrs = model_conv.fc.in_features
model_conv.fc = nn.Linear(num_ftrs, 2)

model_conv = model_conv.to(device)

criterion = nn.CrossEntropyLoss()

# Observe that only parameters of final layer are being optimized as
# opposed to before.
optimizer_conv = optim.SGD(model_conv.fc.parameters(), lr=0.001, momentum=0.9)

# Decay LR by a factor of 0.1 every 7 epochs
exp_lr_scheduler = lr_scheduler.StepLR(optimizer_conv, step_size=7, gamma=0.1)
# 训练并评估
model_conv = train_model(model_conv, criterion, optimizer_conv, exp_lr_scheduler, num_epochs=25)
#输出
Epoch 0/24
----------
train Loss: 0.5447 Acc: 0.7459
val Loss: 0.2617 Acc: 0.9085
...
Epoch 24/24
----------
train Loss: 0.3522 Acc: 0.8443
val Loss: 0.1955 Acc: 0.9412

Training complete in 28m 3s
Best val Acc: 0.947712