论文地址:http://arxiv.org/abs/1602.07261
2014年,GoogLeNet和VGG是当年ImageNet挑战赛(ILSVRC14)的双雄,GoogLeNet获得了第一名、VGG获得了第二名,这两类模型结构的共同特点是层次更深了。VGG继承了LeNet以及AlexNet的一些框架结构,而GoogLeNet则做了更加大胆的网络结构尝试,虽然深度只有22层,但大小却比AlexNet和VGG小很多,GoogleNet参数为500万个,AlexNet参数个数是GoogleNet的12倍,VGGNet参数又是AlexNet的3倍,因此在内存或计算资源有限时,GoogleNet是比较好的选择;从模型结果来看,GoogLeNet的性能却更加优越。
那么,GoogLeNet是如何进一步提升性能的呢?
一般来说,提升网络性能最直接的办法就是增加网络深度和宽度,深度指网络层次数量、宽度指神经元数量。但这种方式存在以下问题:
(1)参数太多,如果训练数据集有限,很容易产生过拟合;
(2)网络越大、参数越多,计算复杂度越大,难以应用;
(3)网络越深,容易出现梯度弥散问题(梯度越往后穿越容易消失),难以优化模型。
所以,有人调侃“深度学习”其实是“深度调参”。
解决这些问题的方法当然就是在增加网络深度和宽度的同时减少参数,为了减少参数,自然就想到将全连接变成稀疏连接。但是在实现上,全连接变成稀疏连接后实际计算量并不会有质的提升,因为大部分硬件是针对密集矩阵计算优化的,稀疏矩阵虽然数据量少,但是计算所消耗的时间却很难减少。
网络中的亮点:
注意:每个分支所得的特征矩阵高和宽必须相同
左图呢,是论文中提出的inception原始结构,右图是inception加上降维功能的结构。
先看左图,inception结构一共有4个分支,也就是说我们的输入的特征矩阵并行的通过这四个分支得到四个输出,然后在将这四个输出在深度维度(channel维度)进行拼接得到我们的最终输出(注意,为了让四个分支的输出能够在深度方向进行拼接,必须保证四个分支输出的特征矩阵高度和宽度都相同)。
再看右图,对比左图,就是在分支2,3,4上加入了卷积核大小为1x1的卷积层,目的是为了降维,减少模型训练参数,减少计算量。
1x1卷积核降维
可以看到图1中有多个黄色的1x1卷积模块,这样的卷积有什么用处呢?
作用1:在相同尺寸的感受野中叠加更多的卷积,能提取到更丰富的特征。
上图左侧是是传统的卷积层结构(线性卷积),在一个尺度上只有一次卷积;右图是Network in Network结构(NIN结构),先进行一次普通的卷积(比如3x3),紧跟再进行一次1x1的卷积,对于某个像素点来说1x1卷积等效于该像素点在所有特征上进行一次全连接的计算,所以右侧图的1x1卷积画成了全连接层的形式,需要注意的是NIN结构中无论是第一个3x3卷积还是新增的1x1卷积,后面都紧跟着激活函数(比如relu)。将两个卷积串联,就能组合出更多的非线性特征。
举个例子,假设第1个3x3卷积+激活函数近似于f1(x)=ax2+bx+c,第二个1x1卷积+激活函数近似于f2(x)=mx2+nx+q,那f1(x)和f2(f1(x))比哪个非线性更强,更能模拟非线性的特征?答案是显而易见的。NIN的结构和传统的神经网络中多层的结构有些类似,后者的多层是跨越了不同尺寸的感受野(通过层与层中间加pool层),从而在更高尺度上提取出特征;NIN结构是在同一个尺度上的多层(中间没有pool层),从而在相同的感受野范围能提取更强的非线性。
作用2:使用1x1卷积进行降维,降低了计算复杂度。
同样是对一个深度为512的特征矩阵使用65个大小为5x5的卷积核进行卷积,不使用1x1卷积核进行降维话一共需要819200个参数,如果使用1x1卷积核进行降维一共需要50688个参数,明显少了很多。
每个卷积层的卷积核个数如何确定呢,下面是原论文中给出的参数列表,对于我们搭建的Inception模块,所需要使用到参数有#1x1, #3x3reduce, #3x3, #5x5reduce, #5x5, poolproj,这6个参数,分别对应着所使用的卷积核个数。
下面这幅图是我将Inception模块所使用到的参数信息标注在每个分支上,其中#1x1对应着分支1上1x1的卷积核个数,#3x3reduce对应着分支2上1x1的卷积核个数,#3x3对应着分支2上3x3的卷积核个数,#5x5reduce对应着分支3上1x1的卷积核个数,#5x5对应着分支3上5x5的卷积核个数,poolproj对应着分支4上1x1的卷积核个数。
上图可以看到对输入做了4个分支,分别用不同尺寸的filter进行卷积或池化,最后再在特征维度上拼接到一起。这种全新的结构有什么好处呢?Szegedy从多个角度进行了解释:
解释1:在直观感觉上在多个尺度上同时进行卷积,能提取到不同尺度的特征。特征更为丰富也意味着最后分类判断时更加准确。
解释2:利用稀疏矩阵分解成密集矩阵计算的原理来加快收敛速度。
举个例子下图左侧是个稀疏矩阵(很多元素都为0,不均匀分布在矩阵中),和一个2x2的矩阵进行卷积,需要对稀疏矩阵中的每一个元素进行计算;如果像右图那样把稀疏矩阵分解成2个子密集矩阵,再和2x2矩阵进行卷积,稀疏矩阵中0较多的区域就可以不用计算,计算量就大大降低。这个原理应用到inception上就是要在特征维度上进行分解!
传统的卷积层的输入数据只和一种尺度(比如3x3)的卷积核进行卷积,输出固定维度(比如256个特征)的数据,所有256个输出特征基本上是均匀分布在3x3尺度范围上,这可以理解成输出了一个稀疏分布的特征集;而inception模块在多个尺度上提取特征(比如1x1,3x3,5x5),输出的256个特征就不再是均匀分布,而是相关性强的特征聚集在一起(比如1x1的的96个特征聚集在一起,3x3的96个特征聚集在一起,5x5的64个特征聚集在一起),这可以理解成多个密集分布的子特征集。这样的特征集中因为相关性较强的特征聚集在了一起,不相关的非关键特征就被弱化,同样是输出256个特征,inception方法输出的特征“冗余”的信息较少。用这样的“纯”的特征集层层传递最后作为反向计算的输入,自然收敛的速度更快。
解释3:Hebbin赫布原理。
Hebbin原理是神经科学上的一个理论,解释了在学习的过程中脑中的神经元所发生的变化,用一句话概括就是fire togethter, wire together。赫布认为“两个神经元或者神经元系统,如果总是同时兴奋,就会形成一种‘组合’,其中一个神经元的兴奋会促进另一个的兴奋”。比如狗看到肉会流口水,反复刺激后,脑中识别肉的神经元会和掌管唾液分泌的神经元会相互促进,“缠绕”在一起,以后再看到肉就会更快流出口水。用在inception结构中就是要把相关性强的特征汇聚到一起。这有点类似上面的解释2,把1x1,3x3,5x5的特征分开。因为训练收敛的最终目的就是要提取出独立的特征,所以预先把相关性强的特征汇聚,就能起到加速收敛的作用。
在inception模块中有一个分支使用了max pooling,作者认为pooling也能起到提取特征的作用,所以也加入模块中。注意这个pooling的stride=1,pooling后没有减少数据的尺寸
辅助分类器结构
网络中的两个辅助分类器结构是一模一样的,如下图所示
着两个辅助分类器的输入分别来自Inception(4a)和Inception(4d)。
[v1] Going Deeper withConvolutions, 6.67% test error,2014.9
论文地址:http://arxiv.org/abs/1409.4842
[v2] Batch Normalization:Accelerating Deep Network Training by Reducing Internal Covariate Shift, 4.8% test error,2015.2
论文地址:http://arxiv.org/abs/1502.03167
[v3] Rethinking theInception Architecture for Computer Vision, 3.5%test error,2015.12
论文地址:http://arxiv.org/abs/1512.00567
[v4] Inception-v4,Inception-ResNet and the Impact of Residual Connections on Learning, 3.08% test error,2016.2
注:本次训练集下载在AlexNet博客有详细解说:https://blog.csdn.net/weixin_44023658/article/details/105798326
#model.py
import torch.nn as nn
import torch
import torch.nn.functional as F
class GoogLeNet(nn.Module):
def __init__(self, num_classes=1000, aux_logits=True, init_weights=False):
super(GoogLeNet, self).__init__()
self.aux_logits = aux_logits
self.conv1 = BasicConv2d(3, 64, kernel_size=7, stride=2, padding=3)
self.maxpool1 = nn.MaxPool2d(3, stride=2, ceil_mode=True)
self.conv2 = BasicConv2d(64, 64, kernel_size=1)
self.conv3 = BasicConv2d(64, 192, kernel_size=3, padding=1)
self.maxpool2 = nn.MaxPool2d(3, stride=2, ceil_mode=True)
self.inception3a = Inception(192, 64, 96, 128, 16, 32, 32)
self.inception3b = Inception(256, 128, 128, 192, 32, 96, 64)
self.maxpool3 = nn.MaxPool2d(3, stride=2, ceil_mode=True)
self.inception4a = Inception(480, 192, 96, 208, 16, 48, 64)
self.inception4b = Inception(512, 160, 112, 224, 24, 64, 64)
self.inception4c = Inception(512, 128, 128, 256, 24, 64, 64)
self.inception4d = Inception(512, 112, 144, 288, 32, 64, 64)
self.inception4e = Inception(528, 256, 160, 320, 32, 128, 128)
self.maxpool4 = nn.MaxPool2d(3, stride=2, ceil_mode=True)
self.inception5a = Inception(832, 256, 160, 320, 32, 128, 128)
self.inception5b = Inception(832, 384, 192, 384, 48, 128, 128)
if self.aux_logits:
self.aux1 = InceptionAux(512, num_classes)
self.aux2 = InceptionAux(528, num_classes)
self.avgpool = nn.AdaptiveAvgPool2d((1, 1))
self.dropout = nn.Dropout(0.4)
self.fc = nn.Linear(1024, num_classes)
if init_weights:
self._initialize_weights()
def forward(self, x):
# N x 3 x 224 x 224
x = self.conv1(x)
# N x 64 x 112 x 112
x = self.maxpool1(x)
# N x 64 x 56 x 56
x = self.conv2(x)
# N x 64 x 56 x 56
x = self.conv3(x)
# N x 192 x 56 x 56
x = self.maxpool2(x)
# N x 192 x 28 x 28
x = self.inception3a(x)
# N x 256 x 28 x 28
x = self.inception3b(x)
# N x 480 x 28 x 28
x = self.maxpool3(x)
# N x 480 x 14 x 14
x = self.inception4a(x)
# N x 512 x 14 x 14
if self.training and self.aux_logits: # eval model lose this layer
aux1 = self.aux1(x)
x = self.inception4b(x)
# N x 512 x 14 x 14
x = self.inception4c(x)
# N x 512 x 14 x 14
x = self.inception4d(x)
# N x 528 x 14 x 14
if self.training and self.aux_logits: # eval model lose this layer
aux2 = self.aux2(x)
x = self.inception4e(x)
# N x 832 x 14 x 14
x = self.maxpool4(x)
# N x 832 x 7 x 7
x = self.inception5a(x)
# N x 832 x 7 x 7
x = self.inception5b(x)
# N x 1024 x 7 x 7
x = self.avgpool(x)
# N x 1024 x 1 x 1
x = torch.flatten(x, 1)
# N x 1024
x = self.dropout(x)
x = self.fc(x)
# N x 1000 (num_classes)
if self.training and self.aux_logits: # eval model lose this layer
return x, aux2, aux1
return x
def _initialize_weights(self):
for m in self.modules():
if isinstance(m, nn.Conv2d):
nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')
if m.bias is not None:
nn.init.constant_(m.bias, 0)
elif isinstance(m, nn.Linear):
nn.init.normal_(m.weight, 0, 0.01)
nn.init.constant_(m.bias, 0)
#inception结构
class Inception(nn.Module):
def __init__(self, in_channels, ch1x1, ch3x3red, ch3x3, ch5x5red, ch5x5, pool_proj):
super(Inception, self).__init__()
self.branch1 = BasicConv2d(in_channels, ch1x1, kernel_size=1)
self.branch2 = nn.Sequential(
BasicConv2d(in_channels, ch3x3red, kernel_size=1),
BasicConv2d(ch3x3red, ch3x3, kernel_size=3, padding=1) # 保证输出大小等于输入大小
)
self.branch3 = nn.Sequential(
BasicConv2d(in_channels, ch5x5red, kernel_size=1),
BasicConv2d(ch5x5red, ch5x5, kernel_size=5, padding=2) # 保证输出大小等于输入大小
)
self.branch4 = nn.Sequential(
nn.MaxPool2d(kernel_size=3, stride=1, padding=1),
BasicConv2d(in_channels, pool_proj, kernel_size=1)
)
def forward(self, x):
branch1 = self.branch1(x)
branch2 = self.branch2(x)
branch3 = self.branch3(x)
branch4 = self.branch4(x)
outputs = [branch1, branch2, branch3, branch4]
return torch.cat(outputs, 1)
#辅助分类器
class InceptionAux(nn.Module):
def __init__(self, in_channels, num_classes):
super(InceptionAux, self).__init__()
self.averagePool = nn.AvgPool2d(kernel_size=5, stride=3)
self.conv = BasicConv2d(in_channels, 128, kernel_size=1) # output[batch, 128, 4, 4]
self.fc1 = nn.Linear(2048, 1024)
self.fc2 = nn.Linear(1024, num_classes)
def forward(self, x):
# aux1: N x 512 x 14 x 14, aux2: N x 528 x 14 x 14
x = self.averagePool(x)
# aux1: N x 512 x 4 x 4, aux2: N x 528 x 4 x 4
x = self.conv(x)
# N x 128 x 4 x 4
x = torch.flatten(x, 1)
x = F.dropout(x, 0.5, training=self.training)
# N x 2048
x = F.relu(self.fc1(x), inplace=True)
x = F.dropout(x, 0.5, training=self.training)
# N x 1024
x = self.fc2(x)
# N x num_classes
return x
class BasicConv2d(nn.Module):
def __init__(self, in_channels, out_channels, **kwargs):
super(BasicConv2d, self).__init__()
self.conv = nn.Conv2d(in_channels, out_channels, **kwargs)
self.relu = nn.ReLU(inplace=True)
def forward(self, x):
x = self.conv(x)
x = self.relu(x)
return x
#train.py
import torch
import torch.nn as nn
from torchvision import transforms, datasets
import torchvision
import json
import matplotlib.pyplot as plt
import os
import torch.optim as optim
from model import GoogLeNet
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
print(device)
data_transform = {
"train": transforms.Compose([transforms.RandomResizedCrop(224),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))]),
"val": transforms.Compose([transforms.Resize((224, 224)),
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])}
#data_root = os.path.abspath(os.path.join(os.getcwd(), "../..")) # get data root path
data_root = os.getcwd()
image_path = data_root + "/flower_data/" # flower data set path
train_dataset = datasets.ImageFolder(root=image_path + "train",
transform=data_transform["train"])
train_num = len(train_dataset)
# {'daisy':0, 'dandelion':1, 'roses':2, 'sunflower':3, 'tulips':4}
flower_list = train_dataset.class_to_idx
cla_dict = dict((val, key) for key, val in flower_list.items())
# write dict into json file
json_str = json.dumps(cla_dict, indent=4)
with open('class_indices.json', 'w') as json_file:
json_file.write(json_str)
batch_size = 32
train_loader = torch.utils.data.DataLoader(train_dataset,
batch_size=batch_size, shuffle=True,
num_workers=0)
validate_dataset = datasets.ImageFolder(root=image_path + "val",
transform=data_transform["val"])
val_num = len(validate_dataset)
validate_loader = torch.utils.data.DataLoader(validate_dataset,
batch_size=batch_size, shuffle=False,
num_workers=0)
# test_data_iter = iter(validate_loader)
# test_image, test_label = test_data_iter.next()
# net = torchvision.models.googlenet(num_classes=5)
# model_dict = net.state_dict()
# pretrain_model = torch.load("googlenet.pth")
# del_list = ["aux1.fc2.weight", "aux1.fc2.bias",
# "aux2.fc2.weight", "aux2.fc2.bias",
# "fc.weight", "fc.bias"]
# pretrain_dict = {k: v for k, v in pretrain_model.items() if k not in del_list}
# model_dict.update(pretrain_dict)
# net.load_state_dict(model_dict)
net = GoogLeNet(num_classes=5, aux_logits=True, init_weights=True)
net.to(device)
loss_function = nn.CrossEntropyLoss()
optimizer = optim.Adam(net.parameters(), lr=0.0003)
best_acc = 0.0
save_path = './googleNet.pth'
for epoch in range(30):
# train
net.train()
running_loss = 0.0
for step, data in enumerate(train_loader, start=0):
images, labels = data
optimizer.zero_grad()
logits, aux_logits2, aux_logits1 = net(images.to(device))
loss0 = loss_function(logits, labels.to(device))
loss1 = loss_function(aux_logits1, labels.to(device))
loss2 = loss_function(aux_logits2, labels.to(device))
loss = loss0 + loss1 * 0.3 + loss2 * 0.3
loss.backward()
optimizer.step()
# print statistics
running_loss += loss.item()
# print train process
rate = (step + 1) / len(train_loader)
a = "*" * int(rate * 50)
b = "." * int((1 - rate) * 50)
print("\rtrain loss: {:^3.0f}%[{}->{}]{:.3f}".format(int(rate * 100), a, b, loss), end="")
print()
# validate
net.eval()
acc = 0.0 # accumulate accurate number / epoch
with torch.no_grad():
for val_data in validate_loader:
val_images, val_labels = val_data
outputs = net(val_images.to(device)) # eval model only have last output layer
predict_y = torch.max(outputs, dim=1)[1]
acc += (predict_y == val_labels.to(device)).sum().item()
val_accurate = acc / val_num
if val_accurate > best_acc:
best_acc = val_accurate
torch.save(net.state_dict(), save_path)
print('[epoch %d] train_loss: %.3f test_accuracy: %.3f' %
(epoch + 1, running_loss / step, val_accurate))
print('Finished Training')
#predict.py
import torch
from model import GoogLeNet
from PIL import Image
from torchvision import transforms
import matplotlib.pyplot as plt
import json
data_transform = transforms.Compose(
[transforms.Resize((224, 224)),
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])
# load image
img = Image.open("../tulip.jpg")
plt.imshow(img)
# [N, C, H, W]
img = data_transform(img)
# expand batch dimension
img = torch.unsqueeze(img, dim=0)
# read class_indict
try:
json_file = open('./class_indices.json', 'r')
class_indict = json.load(json_file)
except Exception as e:
print(e)
exit(-1)
# create model
model = GoogLeNet(num_classes=5, aux_logits=False)
# load model weights
model_weight_path = "./googleNet.pth"
missing_keys, unexpected_keys = model.load_state_dict(torch.load(model_weight_path), strict=False)
model.eval()
with torch.no_grad():
# predict class
output = torch.squeeze(model(img))
predict = torch.softmax(output, dim=0)
predict_cla = torch.argmax(predict).numpy()
print(class_indict[str(predict_cla)])
plt.show()
参考自:
太阳花的小绿豆