打卡
目录
打卡
图像分类
ResNet网络介绍
数据集准备与加载
可视化部分数据集
残差网络构建
Building Block 结构
代码实现
Bottleneck结构
代码实现
构建ResNet50网络
代码定义
模型训练与评估
可视化模型预测
重点:通过网络层数加深,感知CNN和RNN的差别之处,感知不同残差网络结构的差异(Building Block 和 Bottleneck)。
图像分类
图像分类属于有监督学习类别。本案例主要用ResNet50网络对CIFAR-10数据集进行分类。
ResNet网络介绍
传统的 CNN 是将一系列的卷积层和池化层堆叠,但当网络堆叠到一定深度时,就会出现退化问题,即 网络层数的加深并不能改善网络的训练/测试误差。
ResNet网络提出了残差网络结构 (Residual Network) 来减轻退化问题,使用ResNet网络可以实现搭建较深的网络结构(突破1000层)。研究表明ResNet网络层数越深,其训练误差和测试误差越小 (见 ResNet 论文)。
数据集准备与加载
CIFAR-10数据集 共有60000张32*32的彩色图像,分为10个类别,每类有6000张图,数据集一共有50000张训练图片和10000张评估图片。
如下代码,目前仅支持解析二进制版本的CIFAR-10文件(CIFAR-10 binary version)。
from download import downloadurl = "https://mindspore-website.obs.cn-north-4.myhuaweicloud.com/notebook/datasets/cifar-10-binary.tar.gz"download(url, "./datasets-cifar10-bin", kind="tar.gz", replace=True)
如下代码,用mindspore.dataset.Cifar10Dataset
接口来加载数据集,并进行相关图像增强操作。
import mindspore as ms
import mindspore.dataset as ds
import mindspore.dataset.vision as vision
import mindspore.dataset.transforms as transforms
from mindspore import dtype as mstypedata_dir = "./datasets-cifar10-bin/cifar-10-batches-bin" # 数据集根目录
batch_size = 256 # 批量大小
image_size = 32 # 训练图像空间大小
workers = 4 # 并行线程个数
num_classes = 10 # 分类数量def create_dataset_cifar10(dataset_dir, usage, resize, batch_size, workers):data_set = ds.Cifar10Dataset(dataset_dir=dataset_dir,usage=usage,num_parallel_workers=workers,shuffle=True)trans = []if usage == "train":trans += [vision.RandomCrop((32, 32), (4, 4, 4, 4)),vision.RandomHorizontalFlip(prob=0.5)]trans += [vision.Resize(resize),vision.Rescale(1.0 / 255.0, 0.0),vision.Normalize([0.4914, 0.4822, 0.4465], [0.2023, 0.1994, 0.2010]),vision.HWC2CHW()]target_trans = transforms.TypeCast(mstype.int32)# 数据映射操作data_set = data_set.map(operations=trans,input_columns='image',num_parallel_workers=workers)data_set = data_set.map(operations=target_trans,input_columns='label',num_parallel_workers=workers)# 批量操作data_set = data_set.batch(batch_size)return data_set# 获取处理后的训练与测试数据集dataset_train = create_dataset_cifar10(dataset_dir=data_dir,usage="train",resize=image_size,batch_size=batch_size,workers=workers)
step_size_train = dataset_train.get_dataset_size()dataset_val = create_dataset_cifar10(dataset_dir=data_dir,usage="test",resize=image_size,batch_size=batch_size,workers=workers)
step_size_val = dataset_val.get_dataset_size()
可视化部分数据集
import matplotlib.pyplot as plt
import numpy as npdata_iter = next(dataset_train.create_dict_iterator())images = data_iter["image"].asnumpy()
labels = data_iter["label"].asnumpy()
print(f"Image shape: {images.shape}, Label shape: {labels.shape}")# 训练数据集中,前六张图片所对应的标签
print(f"Labels: {labels[:6]}")classes = []with open(data_dir + "/batches.meta.txt", "r") as f:for line in f:line = line.rstrip()if line:classes.append(line)# 训练数据集的前六张图片
plt.figure()
for i in range(6):plt.subplot(2, 3, i + 1)image_trans = np.transpose(images[i], (1, 2, 0))mean = np.array([0.4914, 0.4822, 0.4465])std = np.array([0.2023, 0.1994, 0.2010])image_trans = std * image_trans + meanimage_trans = np.clip(image_trans, 0, 1)plt.title(f"{classes[labels[i]]}")plt.imshow(image_trans)plt.axis("off")
plt.show()
残差网络构建
残差网络结构图如下图所示。
- 残差网络由两个分支构成:一个主分支,一个shortcuts(图中弧线表示)。两条分支输出的特征矩阵相加得到 𝐹(𝑥) + 𝑥,通过Relu 激活函数后即为残差网络最后的输出。
- 残差网络结构主要有两种,一种是Building Block,适用于较浅的ResNet网络,如ResNet18和ResNet34;另一种是Bottleneck,适用于层数较深的ResNet网络,如ResNet50、ResNet101和ResNet152。
Building Block 结构
Building Block 结构的參差网络主分支有两层卷积网络结构:
- 第一层网络以输入channel为64为例,首先通过一个3×3的卷积层,然后通过Batch Normalization层,最后通过Relu激活函数层,输出channel为64;
- 第二层网络类似,只是少了Relu激活函数层。
注意,主分支与shortcuts输出的特征矩阵相加时,需要保证主分支与shortcuts输出的特征矩阵shape相同。
代码实现
如下,实现较为朴素,继承 mindspore.nn.Cell 基层网络,构造函数中初始化实现了网络参数定义,construct 中构建了两层网络主分支结构并进行了残差计算。
from typing import Type, Union, List, Optional
import mindspore.nn as nn
from mindspore.common.initializer import Normal# 初始化卷积层与BatchNorm的参数
weight_init = Normal(mean=0, sigma=0.02)
gamma_init = Normal(mean=1, sigma=0.02)class ResidualBlockBase(nn.Cell):expansion: int = 1 # 最后一个卷积核数量与第一个卷积核数量相等def __init__(self, in_channel: int, out_channel: int,stride: int = 1, norm: Optional[nn.Cell] = None,down_sample: Optional[nn.Cell] = None) -> None:super(ResidualBlockBase, self).__init__()if not norm:self.norm = nn.BatchNorm2d(out_channel)else:self.norm = normself.conv1 = nn.Conv2d(in_channel, out_channel,kernel_size=3, stride=stride,weight_init=weight_init)self.conv2 = nn.Conv2d(in_channel, out_channel,kernel_size=3, weight_init=weight_init)self.relu = nn.ReLU()self.down_sample = down_sampledef construct(self, x):"""ResidualBlockBase construct."""identity = x # shortcuts分支out = self.conv1(x) # 主分支第一层:3*3卷积层out = self.norm(out)out = self.relu(out)out = self.conv2(out) # 主分支第二层:3*3卷积层out = self.norm(out)if self.down_sample is not None:identity = self.down_sample(x)out += identity # 输出为主分支与shortcuts之和out = self.relu(out)return out
Bottleneck结构
相比 Building Block 结构,同等情况下,Bottleneck 结构的参数数量更少,更适合层数较深的网络,ResNet50使用的残差结构就是Bottleneck。
该结构的主分支有三层卷积结构,分别为1×1 的卷积层、3×3卷积层和1×1 的卷积层,其中两层1×1的卷积层起降维、升维的作用。
- 第一层网络以输入channel为256为例,首先通过数量为64,大小为1×1 卷积核进行降维,然后通过Batch Normalization层,最后通过Relu激活函数层,其输出channel为64;
- 第二层网络通过数量为64,大小为3×3 的卷积核提取特征,然后通过Batch Normalization层,最后通过Relu激活函数层,其输出channel为64;
- 第三层通过数量为256,大小1×1 的卷积核进行升维,然后通过Batch Normalization层,其输出channel为256。
同样地,主分支与shortcuts输出的特征矩阵相加时,需要保证主分支与shortcuts输出的特征矩阵shape相同。
代码实现
class ResidualBlock(nn.Cell):expansion = 4 # 最后一个卷积核的数量是第一个卷积核数量的4倍def __init__(self, in_channel: int, out_channel: int,stride: int = 1, down_sample: Optional[nn.Cell] = None) -> None:super(ResidualBlock, self).__init__()self.conv1 = nn.Conv2d(in_channel, out_channel,kernel_size=1, weight_init=weight_init)self.norm1 = nn.BatchNorm2d(out_channel)self.conv2 = nn.Conv2d(out_channel, out_channel,kernel_size=3, stride=stride,weight_init=weight_init)self.norm2 = nn.BatchNorm2d(out_channel)self.conv3 = nn.Conv2d(out_channel, out_channel * self.expansion,kernel_size=1, weight_init=weight_init)self.norm3 = nn.BatchNorm2d(out_channel * self.expansion)self.relu = nn.ReLU()self.down_sample = down_sampledef construct(self, x):identity = x # shortscuts分支out = self.conv1(x) # 主分支第一层:1*1卷积层out = self.norm1(out)out = self.relu(out)out = self.conv2(out) # 主分支第二层:3*3卷积层out = self.norm2(out)out = self.relu(out)out = self.conv3(out) # 主分支第三层:1*1卷积层out = self.norm3(out)if self.down_sample is not None:identity = self.down_sample(x)out += identity # 输出为主分支与shortcuts之和out = self.relu(out)return out
构建ResNet50网络
ResNet50 网络结构图如下。
以输入彩色图像224×224 为例,首先通过数量64,卷积核大小为7×7,stride为2的卷积层conv1,该层输出图片大小为112×112 ,输出channel为64;然后通过一个3×3 的最大下采样池化层,该层输出图片大小为56×56 ,输出channel为64;再堆叠4个残差网络块(conv2_x、conv3_x、conv4_x和conv5_x),此时输出图片大小为7×7 ,输出channel为2048;最后通过一个平均池化层、全连接层和softmax,得到分类概率。
如下代码,make_layer
实现残差块的构建。
def make_layer( last_out_channel, block: Type[Union[ResidualBlockBase, ResidualBlock]],channel: int, block_nums: int, stride: int = 1):"""last_out_channel:上一个残差网络输出的通道数。block:残差网络的类别,分别为ResidualBlockBase和ResidualBlock。channel:残差网络输入的通道数。block_nums:残差网络块堆叠的个数。stride:卷积移动的步幅。"""down_sample = None # shortcuts分支if stride != 1 or last_out_channel != channel * block.expansion:down_sample = nn.SequentialCell([nn.Conv2d(last_out_channel, channel * block.expansion,kernel_size=1, stride=stride, weight_init=weight_init),nn.BatchNorm2d(channel * block.expansion, gamma_init=gamma_init)])layers = []layers.append(block(last_out_channel, channel, stride=stride, down_sample=down_sample))in_channel = channel * block.expansion# 堆叠残差网络for _ in range(1, block_nums):layers.append(block(in_channel, channel))return nn.SequentialCell(layers)
代码定义
ResNet50网络共有5个卷积结构,一个平均池化层,一个全连接层。以 CIFAR-10 数据集为例,代码的实现结构说明如下。
- conv1:输入图片大小为32×32 ,输入channel为3。首先经过一个卷积核数量为64,卷积核大小为7×7 ,stride为2的卷积层;然后通过一个Batch Normalization层;最后通过Reul激活函数。该层输出feature map大小为16×16,输出channel为64。
- conv2_x:输入feature map大小为16×16 ,输入channel为64。首先经过一个卷积核大小为3×3 ,stride为2的最大下采样池化操作;然后堆叠3个[1×1,64;3×3,64;1×1,256][1×1,64;3×3,64;1×1,256]结构的Bottleneck。该层输出feature map大小为8×8 ,输出channel为256。
- conv3_x:输入feature map大小为8×8 ,输入channel为256。该层堆叠4个[1×1,128;3×3,128;1×1,512] 结构的Bottleneck。该层输出feature map大小为4×4 ,输出channel为512。
- conv4_x:输入feature map大小为4×4,输入channel为512。该层堆叠6个[1×1,256;3×3,256;1×1,1024]结构的Bottleneck。该层输出feature map大小为2×2,输出channel为1024。
- conv5_x:输入feature map大小为2×2,输入channel为1024。该层堆叠3个[1×1,512;3×3,512;1×1,2048] 结构的Bottleneck。该层输出feature map大小为1×1 ,输出channel为2048。
- average pool & fc:输入channel为2048,输出channel为分类的类别数。
from mindspore import load_checkpoint, load_param_into_netclass ResNet(nn.Cell):def __init__(self, block: Type[Union[ResidualBlockBase, ResidualBlock]],layer_nums: List[int], num_classes: int, input_channel: int) -> None:super(ResNet, self).__init__()self.relu = nn.ReLU()# 第一个卷积层,输入channel为3(彩色图像),输出channel为64self.conv1 = nn.Conv2d(3, 64, kernel_size=7, stride=2, weight_init=weight_init)self.norm = nn.BatchNorm2d(64)# 最大池化层,缩小图片的尺寸self.max_pool = nn.MaxPool2d(kernel_size=3, stride=2, pad_mode='same')# 各个残差网络结构块定义self.layer1 = make_layer(64, block, 64, layer_nums[0])self.layer2 = make_layer(64 * block.expansion, block, 128, layer_nums[1], stride=2)self.layer3 = make_layer(128 * block.expansion, block, 256, layer_nums[2], stride=2)self.layer4 = make_layer(256 * block.expansion, block, 512, layer_nums[3], stride=2)# 平均池化层self.avg_pool = nn.AvgPool2d()# flattern层self.flatten = nn.Flatten()# 全连接层self.fc = nn.Dense(in_channels=input_channel, out_channels=num_classes)def construct(self, x):x = self.conv1(x)x = self.norm(x)x = self.relu(x)x = self.max_pool(x)x = self.layer1(x)x = self.layer2(x)x = self.layer3(x)x = self.layer4(x)x = self.avg_pool(x)x = self.flatten(x)x = self.fc(x)return xdef _resnet(model_url: str, block: Type[Union[ResidualBlockBase, ResidualBlock]],layers: List[int], num_classes: int, pretrained: bool, pretrained_ckpt: str,input_channel: int):model = ResNet(block, layers, num_classes, input_channel)if pretrained:# 加载预训练模型download(url=model_url, path=pretrained_ckpt, replace=True)param_dict = load_checkpoint(pretrained_ckpt)load_param_into_net(model, param_dict)return modeldef resnet50(num_classes: int = 1000, pretrained: bool = False):"""ResNet50模型"""## num_classes:分类的类别数,默认类别数为 1000。## pretrained:下载对应的训练模型,并加载预训练模型中的参数到网络中。resnet50_url = "https://mindspore-website.obs.cn-north-4.myhuaweicloud.com/notebook/models/application/resnet50_224_new.ckpt"resnet50_ckpt = "./LoadPretrainedModel/resnet50_224_new.ckpt"return _resnet(resnet50_url, ResidualBlock, [3, 4, 6, 3], num_classes,pretrained, resnet50_ckpt, 2048)
模型训练与评估
本案例使用 ResNet50预训练模型 进行微调。
通过调用上个小节的 resnet50 函数
构造 ResNet50 模型,并设置pretrained
参数为True,将会自动下载ResNet50预训练模型,并加载预训练模型中的参数到网络中。然后定义优化器和损失函数,逐个epoch打印训练的损失值和评估精度,并保存评估精度最高的ckpt文件(resnet50-best.ckpt)到当前路径的./BestCheckPoint下。
此处我们展示了5个epochs的训练过程,如果想要达到理想的训练效果,建议训练80个epochs。
import os
import mindspore.ops as ops# 定义ResNet50网络
network = resnet50(pretrained=True)# 全连接层输入层的大小
in_channel = network.fc.in_channels
fc = nn.Dense(in_channels=in_channel, out_channels=10)
# 重置全连接层
network.fc = fc# 设置学习率
num_epochs = 5
lr = nn.cosine_decay_lr(min_lr=0.00001, max_lr=0.001, total_step=step_size_train * num_epochs,step_per_epoch=step_size_train, decay_epoch=num_epochs)
# 定义优化器和损失函数
opt = nn.Momentum(params=network.trainable_params(), learning_rate=lr, momentum=0.9)
loss_fn = nn.SoftmaxCrossEntropyWithLogits(sparse=True, reduction='mean')def forward_fn(inputs, targets):logits = network(inputs)loss = loss_fn(logits, targets)return lossgrad_fn = ms.value_and_grad(forward_fn, None, opt.parameters)def train_step(inputs, targets):loss, grads = grad_fn(inputs, targets)opt(grads)return lossdef train(data_loader, epoch):"""模型训练"""losses = []network.set_train(True)for i, (images, labels) in enumerate(data_loader):loss = train_step(images, labels)if i % 100 == 0 or i == step_size_train - 1:print('Epoch: [%3d/%3d], Steps: [%3d/%3d], Train Loss: [%5.3f]' %(epoch + 1, num_epochs, i + 1, step_size_train, loss))losses.append(loss)return sum(losses) / len(losses)def evaluate(data_loader):"""模型验证"""network.set_train(False)correct_num = 0.0 # 预测正确个数total_num = 0.0 # 预测总数for images, labels in data_loader:logits = network(images)pred = logits.argmax(axis=1) # 预测结果correct = ops.equal(pred, labels).reshape((-1, ))correct_num += correct.sum().asnumpy()total_num += correct.shape[0]acc = correct_num / total_num # 准确率return acc# 创建迭代器
data_loader_train = dataset_train.create_tuple_iterator(num_epochs=num_epochs)
data_loader_val = dataset_val.create_tuple_iterator(num_epochs=num_epochs)# 最佳模型存储路径
best_acc = 0
best_ckpt_dir = "./BestCheckpoint"
best_ckpt_path = "./BestCheckpoint/resnet50-best.ckpt"if not os.path.exists(best_ckpt_dir):os.mkdir(best_ckpt_dir)# 开始循环训练
print("Start Training Loop ...")for epoch in range(num_epochs):curr_loss = train(data_loader_train, epoch)curr_acc = evaluate(data_loader_val)print("-" * 50)print("Epoch: [%3d/%3d], Average Train Loss: [%5.3f], Accuracy: [%5.3f]" % (epoch+1, num_epochs, curr_loss, curr_acc))print("-" * 50)# 保存当前预测准确率最高的模型if curr_acc > best_acc:best_acc = curr_accms.save_checkpoint(network, best_ckpt_path)print("=" * 80)
print(f"End of validation the best Accuracy is: {best_acc: 5.3f}, "f"save the best ckpt file in {best_ckpt_path}", flush=True)
可视化模型预测
若预测字体颜色为蓝色表示为预测正确,预测字体颜色为红色则表示预测错误。
import matplotlib.pyplot as pltdef visualize_model(best_ckpt_path, dataset_val):num_class = 10 # 对狼和狗图像进行二分类net = resnet50(num_class)# 加载模型参数param_dict = ms.load_checkpoint(best_ckpt_path)ms.load_param_into_net(net, param_dict)# 加载验证集的数据进行验证data = next(dataset_val.create_dict_iterator())images = data["image"]labels = data["label"]# 预测图像类别output = net(data['image'])pred = np.argmax(output.asnumpy(), axis=1)# 图像分类classes = []with open(data_dir + "/batches.meta.txt", "r") as f:for line in f:line = line.rstrip()if line:classes.append(line)# 显示图像及图像的预测值plt.figure()for i in range(6):plt.subplot(2, 3, i + 1)# 若预测正确,显示为蓝色;若预测错误,显示为红色color = 'blue' if pred[i] == labels.asnumpy()[i] else 'red'plt.title('predict:{}'.format(classes[pred[i]]), color=color)picture_show = np.transpose(images.asnumpy()[i], (1, 2, 0))mean = np.array([0.4914, 0.4822, 0.4465])std = np.array([0.2023, 0.1994, 0.2010])picture_show = std * picture_show + meanpicture_show = np.clip(picture_show, 0, 1)plt.imshow(picture_show)plt.axis('off')plt.show()# 使用测试数据集进行验证
visualize_model(best_ckpt_path=best_ckpt_path, dataset_val=dataset_val)