NA：归一化注意力机制

★★★ 本文源自AlStudio社区精品项目，【点击此处】查看更多精品内容 >>>

NA：归一化注意力机制

摘要

        为了使卷积神经网络（CNN）更高效、更准确，人们提出了各种轻量级的自注意模块。在本文中，我们系统地研究了CNN中最先进的注意模块，发现自注意机制与归一化密切相关。基于这一观察，我们提出了一种新的注意模块，称为归一化注意模块（简称NA模块），几乎无参数。NA模块计算中间特征映射的平均值和标准偏差，并对特征上下文进行归一化处理，这使得CNN模型更易于训练，对信息性特征的响应也更灵敏。
        我们提出的归一化注意模块可以集成到各种基本CNN架构中，并用于许多计算机视觉任务，包括图像识别、对象检测等。在ImageNet和MS COCO基准上的实验结果表明，我们的方法在使用较少参数的情况下优于最先进的方法。

1. NA

        基于自注意机制与归一化的观察本文提出了一种新的注意力模块——归一化注意模块（NA模块），这是一个即插即用模块，与各种基本CNN网络兼容。NA模块使用全局相似性对全局上下文进行建模。然后采用归一化方法对全局上下文进行归一化

1. Context Modeling：使用全局平均池化操作来建模上下文关系
   $$\mathbf{C}=\frac{1}{H\times W}\sum _{h=1}^H\sum _{w=1}^W{\mathbf{X}}_{h,w}$$
   定义上下文相似函数，并计算最终上下文g：
   $$\begin{array}{c}\Phi \left(x_i,\mathbf{C}\right)=-{\Vert x_i-\mathbf{C}\Vert }_1\\ g={\sum }_i^{H\times W}\frac{\Phi \left(x_i,\mathbf{C}\right)}{{\sum }_i^{H\times W}\Phi \left(x_j,\mathbf{C}\right)}\end{array}$$
2. Normalization
   $$\begin{array}{c}\hat{g}=\gamma \frac{g-\mu }{\sigma }+\beta \\ \mu =\frac{1}{C}{\sum }_m^C{g}_m,\sigma ={\left(\frac{1}{C}{\sum }_m^C{\left(g_m-\mu \right)}^2+ϵ\right)}^{\frac{1}{2}}\end{array}$$
3. 最终输出：
   $$\mathbf{Y}=\mathbf{X}\odot \mathrm{sigmoid}(\hat{g})$$

2. 代码复现

2.1 下载并导入所需要的包

!pip install paddlex

%matplotlib inline
import paddle
import paddle.fluid as fluid
import numpy as np
import matplotlib.pyplot as plt
from paddle.vision.datasets import Cifar10
from paddle.vision.transforms import Transpose
from paddle.io import Dataset, DataLoader
from paddle import nn
import paddle.nn.functional as F
import paddle.vision.transforms as transforms
import os
import matplotlib.pyplot as plt
from matplotlib.pyplot import figure
import paddlex
from paddle import ParamAttr

2.2 创建数据集

train_tfm = transforms.Compose([
    transforms.Resize((130, 130)),
    transforms.ColorJitter(brightness=0.2,contrast=0.2, saturation=0.2),
    paddlex.transforms.MixupImage(),
    transforms.RandomResizedCrop(128, scale=(0.6, 1.0)),
    transforms.RandomHorizontalFlip(0.5),
    transforms.RandomRotation(20),
    transforms.ToTensor(),
    transforms.Normalize(mean=(0.485, 0.456, 0.406), std=(0.229, 0.224, 0.225)),
])

test_tfm = transforms.Compose([
    transforms.Resize((128, 128)),
    transforms.ToTensor(),
    transforms.Normalize(mean=(0.485, 0.456, 0.406), std=(0.229, 0.224, 0.225)),
])

paddle.vision.set_image_backend('cv2')
# 使用Cifar10数据集
train_dataset = Cifar10(data_file='data/data152754/cifar-10-python.tar.gz', mode='train', transform = train_tfm, )
val_dataset = Cifar10(data_file='data/data152754/cifar-10-python.tar.gz', mode='test',transform = test_tfm)
print("train_dataset: %d" % len(train_dataset))
print("val_dataset: %d" % len(val_dataset))

train_dataset: 50000
val_dataset: 10000

batch_size=128

train_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True, drop_last=True, num_workers=2)
val_loader = DataLoader(val_dataset, batch_size=batch_size, shuffle=False, drop_last=False, num_workers=2)

2.3 标签平滑

class LabelSmoothingCrossEntropy(nn.Layer):
    def __init__(self, smoothing=0.1):
        super().__init__()
        self.smoothing = smoothing

    def forward(self, pred, target):

        confidence = 1. - self.smoothing
        log_probs = F.log_softmax(pred, axis=-1)
        idx = paddle.stack([paddle.arange(log_probs.shape[0]), target], axis=1)
        nll_loss = paddle.gather_nd(-log_probs, index=idx)
        smooth_loss = paddle.mean(-log_probs, axis=-1)
        loss = confidence * nll_loss + self.smoothing * smooth_loss

        return loss.mean()

2.4 AlexNet-NA

2.4.1 NA

class NA(nn.Layer):
    def __init__(self):
        super().__init__()
        self.gamma = self.create_parameter([1], default_initializer=nn.initializer.Assign(paddle.ones([1])))
        self.beta = self.create_parameter([1], default_initializer=nn.initializer.Assign(paddle.zeros([1])))
        self.softmax = nn.Softmax(axis=2)
        self.sigmoid = nn.Sigmoid()

    def forward(self, x):
        C = paddle.mean(x, axis=[2, 3], keepdim=True)
        
        g = -paddle.abs(x - C)
        b, c, h, w = g.shape
        g = paddle.sum(self.softmax(g.reshape((b, c, -1))), axis=2, keepdim=True).reshape((b, c, 1, 1))

        mean = paddle.mean(g, axis=1, keepdim=True)
        std = paddle.std(g, axis=1, keepdim=True) + 1e-5

        g = self.gamma * (g - mean) / std + self.beta

        out = x * self.sigmoid(g)

        return out
        

此处参数没有统计出来，AlexNet-NA统计的参数是正常的

model = NA()
paddle.summary(model, (1, 64, 224, 224))

W0804 17:19:19.558254   491 gpu_resources.cc:61] Please NOTE: device: 0, GPU Compute Capability: 7.0, Driver API Version: 11.2, Runtime API Version: 10.1
W0804 17:19:19.561859   491 gpu_resources.cc:91] device: 0, cuDNN Version: 7.6.


---------------------------------------------------------------------------
 Layer (type)       Input Shape          Output Shape         Param #    
===========================================================================
   Softmax-1      [[1, 64, 50176]]      [1, 64, 50176]           0       
   Sigmoid-2      [[1, 64, 1, 1]]       [1, 64, 1, 1]            0       
===========================================================================
Total params: 0
Trainable params: 0
Non-trainable params: 0
---------------------------------------------------------------------------
Input size (MB): 12.25
Forward/backward pass size (MB): 24.50
Params size (MB): 0.00
Estimated Total Size (MB): 36.75
---------------------------------------------------------------------------






{'total_params': 0, 'trainable_params': 0}

2.4.2 AlexNet-NA

class AlexNet_NA(nn.Layer):
    def __init__(self,num_classes=10):
        super().__init__()
        self.features=nn.Sequential(
            nn.Conv2D(3,48, kernel_size=11, stride=4, padding=11//2),
            NA(),
            nn.ReLU(),
            nn.MaxPool2D(kernel_size=3,stride=2),
            nn.Conv2D(48,128, kernel_size=5, padding=2),
            NA(),
            nn.ReLU(),
            nn.MaxPool2D(kernel_size=3,stride=2),
            nn.Conv2D(128, 192,kernel_size=3,stride=1,padding=1),
            NA(),
            nn.ReLU(),
            nn.Conv2D(192,192,kernel_size=3,stride=1,padding=1),
            NA(),
            nn.ReLU(),
            nn.Conv2D(192,128,kernel_size=3,stride=1,padding=1),
            NA(),
            nn.ReLU(),
            nn.MaxPool2D(kernel_size=3,stride=2),
        )
        self.classifier=nn.Sequential(
            nn.Linear(3 * 3 * 128,2048),
            nn.ReLU(),
            nn.Dropout(0.5),
            nn.Linear(2048,2048),
            nn.ReLU(),
            nn.Dropout(0.5),
            nn.Linear(2048,num_classes),
        )
 
 
    def forward(self,x):
        x = self.features(x)
        x = paddle.flatten(x, 1)
        x=self.classifier(x)
 
        return x

model = AlexNet_NA(num_classes=10)
paddle.summary(model, (1, 3, 128, 128))

2.5 训练

learning_rate = 0.001
n_epochs = 50
paddle.seed(42)
np.random.seed(42)

work_path = 'work/model'

model = AlexNet_NA(num_classes=10)

criterion = LabelSmoothingCrossEntropy()

scheduler = paddle.optimizer.lr.CosineAnnealingDecay(learning_rate=learning_rate, T_max=50000 // batch_size * n_epochs, verbose=False)
optimizer = paddle.optimizer.Adam(parameters=model.parameters(), learning_rate=scheduler, weight_decay=1e-5)

gate = 0.0
threshold = 0.0
best_acc = 0.0
val_acc = 0.0
loss_record = {'train': {'loss': [], 'iter': []}, 'val': {'loss': [], 'iter': []}}   # for recording loss
acc_record = {'train': {'acc': [], 'iter': []}, 'val': {'acc': [], 'iter': []}}      # for recording accuracy

loss_iter = 0
acc_iter = 0

for epoch in range(n_epochs):
    # ---------- Training ----------
    model.train()
    train_num = 0.0
    train_loss = 0.0

    val_num = 0.0
    val_loss = 0.0
    accuracy_manager = paddle.metric.Accuracy()
    val_accuracy_manager = paddle.metric.Accuracy()
    print("#===epoch: {}, lr={:.10f}===#".format(epoch, optimizer.get_lr()))
    for batch_id, data in enumerate(train_loader):
        x_data, y_data = data
        labels = paddle.unsqueeze(y_data, axis=1)

        logits = model(x_data)

        loss = criterion(logits, y_data)

        acc = paddle.metric.accuracy(logits, labels)
        accuracy_manager.update(acc)
        if batch_id % 10 == 0:
            loss_record['train']['loss'].append(loss.numpy())
            loss_record['train']['iter'].append(loss_iter)
            loss_iter += 1

        loss.backward()

        optimizer.step()
        scheduler.step()
        optimizer.clear_grad()
        
        train_loss += loss
        train_num += len(y_data)

    total_train_loss = (train_loss / train_num) * batch_size
    train_acc = accuracy_manager.accumulate()
    acc_record['train']['acc'].append(train_acc)
    acc_record['train']['iter'].append(acc_iter)
    acc_iter += 1
    # Print the information.
    print("#===epoch: {}, train loss is: {}, train acc is: {:2.2f}%===#".format(epoch, total_train_loss.numpy(), train_acc*100))

    # ---------- Validation ----------
    model.eval()

    for batch_id, data in enumerate(val_loader):

        x_data, y_data = data
        labels = paddle.unsqueeze(y_data, axis=1)
        with paddle.no_grad():
          logits = model(x_data)

        loss = criterion(logits, y_data)

        acc = paddle.metric.accuracy(logits, labels)
        val_accuracy_manager.update(acc)

        val_loss += loss
        val_num += len(y_data)

    total_val_loss = (val_loss / val_num) * batch_size
    loss_record['val']['loss'].append(total_val_loss.numpy())
    loss_record['val']['iter'].append(loss_iter)
    val_acc = val_accuracy_manager.accumulate()
    acc_record['val']['acc'].append(val_acc)
    acc_record['val']['iter'].append(acc_iter)
    
    print("#===epoch: {}, val loss is: {}, val acc is: {:2.2f}%===#".format(epoch, total_val_loss.numpy(), val_acc*100))

    # ===================save====================
    if val_acc > best_acc:
        best_acc = val_acc
        paddle.save(model.state_dict(), os.path.join(work_path, 'best_model.pdparams'))
        paddle.save(optimizer.state_dict(), os.path.join(work_path, 'best_optimizer.pdopt'))

print(best_acc)
paddle.save(model.state_dict(), os.path.join(work_path, 'final_model.pdparams'))
paddle.save(optimizer.state_dict(), os.path.join(work_path, 'final_optimizer.pdopt'))

2.6 实验结果

def plot_learning_curve(record, title='loss', ylabel='CE Loss'):
    ''' Plot learning curve of your CNN '''
    maxtrain = max(map(float, record['train'][title]))
    maxval = max(map(float, record['val'][title]))
    ymax = max(maxtrain, maxval) * 1.1
    mintrain = min(map(float, record['train'][title]))
    minval = min(map(float, record['val'][title]))
    ymin = min(mintrain, minval) * 0.9

    total_steps = len(record['train'][title])
    x_1 = list(map(int, record['train']['iter']))
    x_2 = list(map(int, record['val']['iter']))
    figure(figsize=(10, 6))
    plt.plot(x_1, record['train'][title], c='tab:red', label='train')
    plt.plot(x_2, record['val'][title], c='tab:cyan', label='val')
    plt.ylim(ymin, ymax)
    plt.xlabel('Training steps')
    plt.ylabel(ylabel)
    plt.title('Learning curve of {}'.format(title))
    plt.legend()
    plt.show()

plot_learning_curve(loss_record, title='loss', ylabel='CE Loss')

plot_learning_curve(acc_record, title='acc', ylabel='Accuracy')

import time
work_path = 'work/model'
model = AlexNet_NA(num_classes=10)
model_state_dict = paddle.load(os.path.join(work_path, 'best_model.pdparams'))
model.set_state_dict(model_state_dict)
model.eval()
aa = time.time()
for batch_id, data in enumerate(val_loader):

    x_data, y_data = data
    labels = paddle.unsqueeze(y_data, axis=1)
    with paddle.no_grad():
        logits = model(x_data)
bb = time.time()
print("Throughout:{}".format(int(len(val_dataset)//(bb - aa))))

Throughout:1629

def get_cifar10_labels(labels):  
    """返回CIFAR10数据集的文本标签。"""
    text_labels = [
        'airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'frog',
        'horse', 'ship', 'truck']
    return [text_labels[int(i)] for i in labels]

def show_images(imgs, num_rows, num_cols, pred=None, gt=None, scale=1.5):  
    """Plot a list of images."""
    figsize = (num_cols * scale, num_rows * scale)
    _, axes = plt.subplots(num_rows, num_cols, figsize=figsize)
    axes = axes.flatten()
    for i, (ax, img) in enumerate(zip(axes, imgs)):
        if paddle.is_tensor(img):
            ax.imshow(img.numpy())
        else:
            ax.imshow(img)
        ax.axes.get_xaxis().set_visible(False)
        ax.axes.get_yaxis().set_visible(False)
        if pred or gt:
            ax.set_title("pt: " + pred[i] + "\ngt: " + gt[i])
    return axes

work_path = 'work/model'
X, y = next(iter(DataLoader(val_dataset, batch_size=18)))
model = AlexNet_NA(num_classes=10)
model_state_dict = paddle.load(os.path.join(work_path, 'best_model.pdparams'))
model.set_state_dict(model_state_dict)
model.eval()
logits = model(X)
y_pred = paddle.argmax(logits, -1)
X = paddle.transpose(X, [0, 2, 3, 1])
axes = show_images(X.reshape((18, 128, 128, 3)), 1, 18, pred=get_cifar10_labels(y_pred), gt=get_cifar10_labels(y))
plt.show()

Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).

3. AlexNet

3.1 AlexNet

class AlexNet(nn.Layer):
    def __init__(self,num_classes=10):
        super().__init__()
        self.features=nn.Sequential(
            nn.Conv2D(3,48, kernel_size=11, stride=4, padding=11//2),
            nn.ReLU(),
            nn.MaxPool2D(kernel_size=3,stride=2),
            nn.Conv2D(48,128, kernel_size=5, padding=2),
            nn.ReLU(),
            nn.MaxPool2D(kernel_size=3,stride=2),
            nn.Conv2D(128, 192,kernel_size=3,stride=1,padding=1),
            nn.ReLU(),
            nn.Conv2D(192,192,kernel_size=3,stride=1,padding=1),
            nn.ReLU(),
            nn.Conv2D(192,128,kernel_size=3,stride=1,padding=1),
            nn.ReLU(),
            nn.MaxPool2D(kernel_size=3,stride=2),
        )
        self.classifier=nn.Sequential(
            nn.Linear(3 * 3 * 128,2048),
            nn.ReLU(),
            nn.Dropout(0.5),
            nn.Linear(2048,2048),
            nn.ReLU(),
            nn.Dropout(0.5),
            nn.Linear(2048,num_classes),
        )
 
 
    def forward(self,x):
        x = self.features(x)
        x = paddle.flatten(x, 1)
        x=self.classifier(x)
 
        return x

model = AlexNet(num_classes=10)
paddle.summary(model, (1, 3, 128, 128))

3.2 训练

learning_rate = 0.001
n_epochs = 50
paddle.seed(42)
np.random.seed(42)

work_path = 'work/model1'

model = AlexNet(num_classes=10)

criterion = LabelSmoothingCrossEntropy()

scheduler = paddle.optimizer.lr.CosineAnnealingDecay(learning_rate=learning_rate, T_max=50000 // batch_size * n_epochs, verbose=False)
optimizer = paddle.optimizer.Adam(parameters=model.parameters(), learning_rate=scheduler, weight_decay=1e-5)

gate = 0.0
threshold = 0.0
best_acc = 0.0
val_acc = 0.0
loss_record1 = {'train': {'loss': [], 'iter': []}, 'val': {'loss': [], 'iter': []}}   # for recording loss
acc_record1 = {'train': {'acc': [], 'iter': []}, 'val': {'acc': [], 'iter': []}}      # for recording accuracy

loss_iter = 0
acc_iter = 0

for epoch in range(n_epochs):
    # ---------- Training ----------
    model.train()
    train_num = 0.0
    train_loss = 0.0

    val_num = 0.0
    val_loss = 0.0
    accuracy_manager = paddle.metric.Accuracy()
    val_accuracy_manager = paddle.metric.Accuracy()
    print("#===epoch: {}, lr={:.10f}===#".format(epoch, optimizer.get_lr()))
    for batch_id, data in enumerate(train_loader):
        x_data, y_data = data
        labels = paddle.unsqueeze(y_data, axis=1)

        logits = model(x_data)

        loss = criterion(logits, y_data)

        acc = paddle.metric.accuracy(logits, labels)
        accuracy_manager.update(acc)
        if batch_id % 10 == 0:
            loss_record1['train']['loss'].append(loss.numpy())
            loss_record1['train']['iter'].append(loss_iter)
            loss_iter += 1

        loss.backward()

        optimizer.step()
        scheduler.step()
        optimizer.clear_grad()
        
        train_loss += loss
        train_num += len(y_data)

    total_train_loss = (train_loss / train_num) * batch_size
    train_acc = accuracy_manager.accumulate()
    acc_record1['train']['acc'].append(train_acc)
    acc_record1['train']['iter'].append(acc_iter)
    acc_iter += 1
    # Print the information.
    print("#===epoch: {}, train loss is: {}, train acc is: {:2.2f}%===#".format(epoch, total_train_loss.numpy(), train_acc*100))

    # ---------- Validation ----------
    model.eval()

    for batch_id, data in enumerate(val_loader):

        x_data, y_data = data
        labels = paddle.unsqueeze(y_data, axis=1)
        with paddle.no_grad():
          logits = model(x_data)

        loss = criterion(logits, y_data)

        acc = paddle.metric.accuracy(logits, labels)
        val_accuracy_manager.update(acc)

        val_loss += loss
        val_num += len(y_data)

    total_val_loss = (val_loss / val_num) * batch_size
    loss_record1['val']['loss'].append(total_val_loss.numpy())
    loss_record1['val']['iter'].append(loss_iter)
    val_acc = val_accuracy_manager.accumulate()
    acc_record1['val']['acc'].append(val_acc)
    acc_record1['val']['iter'].append(acc_iter)
    
    print("#===epoch: {}, val loss is: {}, val acc is: {:2.2f}%===#".format(epoch, total_val_loss.numpy(), val_acc*100))

    # ===================save====================
    if val_acc > best_acc:
        best_acc = val_acc
        paddle.save(model.state_dict(), os.path.join(work_path, 'best_model.pdparams'))
        paddle.save(optimizer.state_dict(), os.path.join(work_path, 'best_optimizer.pdopt'))

print(best_acc)
paddle.save(model.state_dict(), os.path.join(work_path, 'final_model.pdparams'))
paddle.save(optimizer.state_dict(), os.path.join(work_path, 'final_optimizer.pdopt'))

3.3 实验结果

plot_learning_curve(loss_record1, title='loss', ylabel='CE Loss')

plot_learning_curve(acc_record1, title='acc', ylabel='Accuracy')

import time
work_path = 'work/model1'
model = AlexNet(num_classes=10)
model_state_dict = paddle.load(os.path.join(work_path, 'best_model.pdparams'))
model.set_state_dict(model_state_dict)
model.eval()
aa = time.time()
for batch_id, data in enumerate(val_loader):

    x_data, y_data = data
    labels = paddle.unsqueeze(y_data, axis=1)
    with paddle.no_grad():
        logits = model(x_data)
bb = time.time()
print("Throughout:{}".format(int(len(val_dataset)//(bb - aa))))

Throughout:1598

work_path = 'work/model1'
X, y = next(iter(DataLoader(val_dataset, batch_size=18)))
model = AlexNet(num_classes=10)
model_state_dict = paddle.load(os.path.join(work_path, 'best_model.pdparams'))
model.set_state_dict(model_state_dict)
model.eval()
logits = model(X)
y_pred = paddle.argmax(logits, -1)
X = paddle.transpose(X, [0, 2, 3, 1])
axes = show_images(X.reshape((18, 128, 128, 3)), 1, 18, pred=get_cifar10_labels(y_pred), gt=get_cifar10_labels(y))
plt.show()

Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).

th RGB data ([0…1] for floats or [0…255] for integers).

4. 对比实验结果

4.1 AlexNet(CIFAR10)

model	Train Acc	Val Acc	parameter
AlexNet w/o NA	0.7700	0.79668	7524042
AlexNet w NA	0.8540	0.84583	7524052

4.2 ResNet18(CIFAR100)

model	Train Acc	Val Acc	parameter
ResNet18 w/o NA	0.9876	0.67346	11237412
ResNet18 w NA	0.9905	0.68434	11237428

实验结果见main-Copy2.ipynb

总结

        论文非常巧妙地将归一化与注意力机制结合，提出了一种新的归一化注意力机制——NA，AlexNet-NA在增加少量参数（+10）的同时大大加快了收敛速度以及精度（+0.04915）;ResNet18在增加少量参数（+16）的同时大大加快了收敛速度以及精度（+0.01088）