>- **🍨 本文为[🔗365天深度学习训练营](https://mp.weixin.qq.com/s/Z9yL_wt7L8aPOr9Lqb1K3w) 中的学习记录博客**
>- **🍖 原作者:[K同学啊](https://mtyjkh.blog.csdn.net/)**
Inception v3由谷歌研究员Christian Szegedy等人在2015年的论文《Rethinking the InceptionArchitecture for Computer Vision》中提出。Inception v3是Inception网络系列的第三个版本,它在ImageNet图像识别竞赛中取得了优异成绩,尤其是在大规模图像识别任务中表现出色Inception v3的主要特点如下:
更深的网络结构:Inception v3比之前的Inception网络结构更深,包含了48层卷积层。这使得网络可1.以提取更多层次的特征,从而在图像识别任务上取得更好的效果。
2使用Factorized Convolutions:Inception v3采用了Factorized Convolutions(分解卷积),将较大的卷积核分解为多个较小的卷积核。这种方法可以降低网络的参数数量,减少计算复杂度,同时保持良好的性能。
使用Batch Normalization:Inception v3在每个卷积层之后都添加了Batch Normalization(BN)这有助于网络的收敛和泛化能力。BN可以减少Internal Covariate Shift(内部协变量偏移)现象,加快训练速度,同时提高模型的鲁棒性。
辅助分类器:Inception v3引入了辅助分类器,可以在网络训练过程中提供额外的梯度信息,帮助网4.络更好地学习特征。辅助分类器位于网络的某个中间层,其输出会与主分类器的输出进行加权融合从而得到最终的预测结果。
基于RMSProp的优化器:Inception v3使用了RMSProp优化器进行训练。相比于传统的随机梯度下降(SGD)方法,RMSProp可以自适应地调整学习率,使得训练过程更加稳定,收敛速度更快。Inception v3在图像分类、物体检测和图像分割等计算机视觉任务中均取得了显著的效果。然而,由于其较大的网络结构和计算复杂度,Inception v3在实际应用中可能需要较高的硬件要求。
运行代码如下:
数据预处理部分
import torch
import torch.nn as nn
import torchvision.transforms as transforms
import torchvision
from torchvision import transforms, datasets
import os,PIL,pathlib,random
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
device
# 导入数据
data_dir = 'D:/研究生课题/深度学习-代码/3-date/weather_photos'
data_dir = pathlib.Path(data_dir)
data_paths = list(data_dir.glob('*'))
classeNames = [str(path).split("\\")[5] for path in data_paths]
classeNames
import matplotlib.pyplot as plt
from PIL import Image
# 指定图像文件夹路径
image_folder = r'D:/研究生课题/深度学习-代码/3-date/weather_photos'
# 获取文件夹中的所有图像文件
image_files = [f for f in os.listdir(image_folder) if f.endswith((".jpg", ".png", ".jpeg"))]
# 创建Matplotlib图像
fig, axes = plt.subplots(3, 8, figsize=(16, 6))
# 使用列表推导式加载和显示图像
for ax, img_file in zip(axes.flat, image_files):
img_path = os.path.join(image_folder, img_file)
img = Image.open(img_path)
ax.imshow(img)
ax.axis('off')
# 显示图像
plt.tight_layout()
plt.show()
total_datadir = 'D:/研究生课题/深度学习-代码/3-date/weather_photos'
# 关于transforms.Compose的更多介绍可以参考:https://blog.csdn.net/qq_38251616/article/details/124878863
train_transforms = transforms.Compose([
transforms.Resize([224, 224]), # 将输入图片resize成统一尺寸
transforms.ToTensor(), # 将PIL Image或numpy.ndarray转换为tensor,并归一化到[0,1]之间
transforms.Normalize( # 标准化处理-->转换为标准正太分布(高斯分布),使模型更容易收敛
mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225]) # 其中 mean=[0.485,0.456,0.406]与std=[0.229,0.224,0.225] 从数据集中随机抽样计算得到的。
])
total_data = datasets.ImageFolder(total_datadir,transform=train_transforms)
train_size = int(0.8 * len(total_data))
test_size = len(total_data) - train_size
train_dataset, test_dataset = torch.utils.data.random_split(total_data, [train_size, test_size])
train_dataset, test_dataset
print(total_data)
# 划分数据集
train_size = int(0.8 * len(total_data))
test_size = len(total_data) - train_size
train_dataset, test_dataset = torch.utils.data.random_split(total_data, [train_size, test_size])
print(train_dataset)
print(test_dataset)
print(train_size,test_size)
batch_size = 32
train_dl = torch.utils.data.DataLoader(train_dataset,
batch_size=batch_size,
shuffle=True,
num_workers=0)
test_dl = torch.utils.data.DataLoader(test_dataset,
batch_size=batch_size,
shuffle=True,
num_workers=0)
for X, y in test_dl:
print("Shape of X [N, C, H, W]: ", X.shape)
print("Shape of y: ", y.shape, y.dtype)
break
搭建模型、训练部分
import torch
import torch.nn as nn
import torch.nn.functional as F
from main import device, test_dl, train_dl
class BasicConv2d(nn.Module):
def __init__(self, in_channel, out_channel, **kwargs):
super(BasicConv2d, self).__init__()
self.conv = nn.Conv2d(in_channel, out_channel, bias=False, **kwargs)
self.norm = nn.BatchNorm2d(out_channel, eps=0.001)
self.relu = nn.ReLU(inplace=True)
def forward(self, x):
x = self.conv(x)
x = self.norm(x)
x = self.relu(x)
return x
class InceptionA(nn.Module):
def __init__(self, in_channels, pool_features):
super(InceptionA, self).__init__()
self.branch1x1 = BasicConv2d(in_channels, 64, kernel_size=1) # 1
self.branch5x5_1 = BasicConv2d(in_channels, 48, kernel_size=1)
self.branch5x5_2 = BasicConv2d(48, 64, kernel_size=5, padding=2)
self.branch3x3dbl_1 = BasicConv2d(in_channels, 64, kernel_size=1)
self.branch3x3dbl_2 = BasicConv2d(64, 96, kernel_size=3, padding=1)
self.branch3x3dbl_3 = BasicConv2d(96, 96, kernel_size=3, padding=1)
self.branch_pool = BasicConv2d(in_channels, pool_features, kernel_size=1)
def forward(self, x):
branch1x1 = self.branch1x1(x)
branch5x5 = self.branch5x5_1(x)
branch5x5 = self.branch5x5_2(branch5x5)
branch3x3dbl = self.branch3x3dbl_1(x)
branch3x3dbl = self.branch3x3dbl_2(branch3x3dbl)
branch3x3dbl = self.branch3x3dbl_3(branch3x3dbl)
branch_pool = F.avg_pool2d(x, kernel_size=3, stride=1, padding=1)
branch_pool = self.branch_pool(branch_pool)
outputs = [branch1x1, branch5x5, branch3x3dbl, branch_pool]
return torch.cat(outputs, 1)
class InceptionB(nn.Module):
def __init__(self, in_channels, channels_7x7):
super(InceptionB, self).__init__()
self.branch1x1 = BasicConv2d(in_channels, 192, kernel_size=1)
c7 = channels_7x7
self.branch7x7_1 = BasicConv2d(in_channels, c7, kernel_size=1)
self.branch7x7_2 = BasicConv2d(c7, c7, kernel_size=(1, 7), padding=(0, 3))
self.branch7x7_3 = BasicConv2d(c7, 192, kernel_size=(7, 1), padding=(3, 0))
self.branch7x7dbl_1 = BasicConv2d(in_channels, c7, kernel_size=1)
self.branch7x7dbl_2 = BasicConv2d(c7, c7, kernel_size=(7, 1), padding=(3, 0))
self.branch7x7dbl_3 = BasicConv2d(c7, c7, kernel_size=(1, 7), padding=(0, 3))
self.branch7x7dbl_4 = BasicConv2d(c7, c7, kernel_size=(7, 1), padding=(3, 0))
self.branch7x7dbl_5 = BasicConv2d(c7, 192, kernel_size=(1, 7), padding=(0, 3))
self.branch_pool = BasicConv2d(in_channels, 192, kernel_size=1)
def forward(self, x):
branch1x1 = self.branch1x1(x)
branch7x7 = self.branch7x7_1(x)
branch7x7 = self.branch7x7_2(branch7x7)
branch7x7 = self.branch7x7_3(branch7x7)
branch7x7dbl = self.branch7x7dbl_1(x)
branch7x7dbl = self.branch7x7dbl_2(branch7x7dbl)
branch7x7dbl = self.branch7x7dbl_3(branch7x7dbl)
branch7x7dbl = self.branch7x7dbl_4(branch7x7dbl)
branch7x7dbl = self.branch7x7dbl_5(branch7x7dbl)
branch_pool = F.avg_pool2d(x, kernel_size=3, stride=1, padding=1)
branch_pool = self.branch_pool(branch_pool)
outputs = [branch1x1, branch7x7, branch7x7dbl, branch_pool]
return torch.cat(outputs, 1)
class InceptionC(nn.Module):
def __init__(self, in_channels):
super(InceptionC, self).__init__()
self.branch1x1 = BasicConv2d(in_channels, 320, kernel_size=1)
self.branch3x3_1 = BasicConv2d(in_channels, 384, kernel_size=1)
self.branch3x3_2a = BasicConv2d(384, 384, kernel_size=(1, 3), padding=(0, 1))
self.branch3x3_2b = BasicConv2d(384, 384, kernel_size=(3, 1), padding=(1, 0))
self.branch3x3dbl_1 = BasicConv2d(in_channels, 448, kernel_size=1)
self.branch3x3dbl_2 = BasicConv2d(448, 384, kernel_size=3, padding=1)
self.branch3x3dbl_3a = BasicConv2d(384, 384, kernel_size=(1, 3), padding=(0, 1))
self.branch3x3dbl_3b = BasicConv2d(384, 384, kernel_size=(3, 1), padding=(1, 0))
self.branch_pool = BasicConv2d(in_channels, 192, kernel_size=1)
def forward(self, x):
branch1x1 = self.branch1x1(x)
branch3x3 = self.branch3x3_1(x)
branch3x3 = [
self.branch3x3_2a(branch3x3),
self.branch3x3_2b(branch3x3),
]
branch3x3 = torch.cat(branch3x3, 1)
branch3x3dbl = self.branch3x3dbl_1(x)
branch3x3dbl = self.branch3x3dbl_2(branch3x3dbl)
branch3x3dbl = [
self.branch3x3dbl_3a(branch3x3dbl),
self.branch3x3dbl_3b(branch3x3dbl),
]
branch3x3dbl = torch.cat(branch3x3dbl, 1)
branch_pool = F.avg_pool2d(x, kernel_size=3, stride=1, padding=1)
branch_pool = self.branch_pool(branch_pool)
outputs = [branch1x1, branch3x3, branch3x3dbl, branch_pool]
return torch.cat(outputs, 1)
class ReductionA(nn.Module):
def __init__(self, in_channels):
super(ReductionA, self).__init__()
self.branch3x3 = BasicConv2d(in_channels, 384, kernel_size=3, stride=2)
self.branch3x3dbl_1 = BasicConv2d(in_channels, 64, kernel_size=1)
self.branch3x3dbl_2 = BasicConv2d(64, 96, kernel_size=3, padding=1)
self.branch3x3dbl_3 = BasicConv2d(96, 96, kernel_size=3, stride=2)
def forward(self, x):
branch3x3 = self.branch3x3(x)
branch3x3dbl = self.branch3x3dbl_1(x)
branch3x3dbl = self.branch3x3dbl_2(branch3x3dbl)
branch3x3dbl = self.branch3x3dbl_3(branch3x3dbl)
branch_pool = F.max_pool2d(x, kernel_size=3, stride=2)
outputs = [branch3x3, branch3x3dbl, branch_pool]
return torch.cat(outputs, 1)
class ReductionB(nn.Module):
def __init__(self, in_channels):
super(ReductionB, self).__init__()
self.branch3x3_1 = BasicConv2d(in_channels, 192, kernel_size=1)
self.branch3x3_2 = BasicConv2d(192, 320, kernel_size=3, stride=2)
self.branch7x7x3_1 = BasicConv2d(in_channels, 192, kernel_size=1)
self.branch7x7x3_2 = BasicConv2d(192, 192, kernel_size=(1, 7), padding=(0, 3))
self.branch7x7x3_3 = BasicConv2d(192, 192, kernel_size=(7, 1), padding=(3, 0))
self.branch7x7x3_4 = BasicConv2d(192, 192, kernel_size=3, stride=2)
def forward(self, x):
branch3x3 = self.branch3x3_1(x)
branch3x3 = self.branch3x3_2(branch3x3)
branch7x7x3 = self.branch7x7x3_1(x)
branch7x7x3 = self.branch7x7x3_2(branch7x7x3)
branch7x7x3 = self.branch7x7x3_3(branch7x7x3)
branch7x7x3 = self.branch7x7x3_4(branch7x7x3)
branch_pool = F.max_pool2d(x, kernel_size=3, stride=2)
outputs = [branch3x3, branch7x7x3, branch_pool]
return torch.cat(outputs, 1)
class InceptionAux(nn.Module):
def __init__(self, in_channels, num_classes):
super(InceptionAux, self).__init__()
self.conv0 = BasicConv2d(in_channels, 128, kernel_size=1)
self.conv1 = BasicConv2d(128, 768, kernel_size=5)
self.conv1.stddev = 0.01
self.fc = nn.Linear(768, num_classes)
self.fc.stddev = 0.001
def forward(self, x):
# 17 x 17 x 768
x = F.avg_pool2d(x, kernel_size=5, stride=3)
# 5 x 5 x 768
x = self.conv0(x)
# 5 x 5 x 128
x = self.conv1(x)
# 1 x 1 x 768
x = x.view(x.size(0), -1)
# 768
x = self.fc(x)
# 1000
return x
import torch.nn.functional as F
class InceptionV3(nn.Module):
def __init__(self, num_classes=1000, aux_logits=False, transform_input=False):
super(InceptionV3, self).__init__()
self.aux_logits = aux_logits
self.transform_input = transform_input
self.Conv2d_1a_3x3 = BasicConv2d(3, 32, kernel_size=3, stride=2)
self.Conv2d_2a_3x3 = BasicConv2d(32, 32, kernel_size=3)
self.Conv2d_2b_3x3 = BasicConv2d(32, 64, kernel_size=3, padding=1)
self.Conv2d_3b_1x1 = BasicConv2d(64, 80, kernel_size=1)
self.Conv2d_4a_3x3 = BasicConv2d(80, 192, kernel_size=3)
self.Mixed_5b = InceptionA(192, pool_features=32)
self.Mixed_5c = InceptionA(256, pool_features=64)
self.Mixed_5d = InceptionA(288, pool_features=64)
self.Mixed_6a = ReductionA(288)
self.Mixed_6b = InceptionB(768, channels_7x7=128)
self.Mixed_6c = InceptionB(768, channels_7x7=160)
self.Mixed_6d = InceptionB(768, channels_7x7=160)
self.Mixed_6e = InceptionB(768, channels_7x7=192)
if aux_logits:
self.AuxLogits = InceptionAux(768, num_classes)
self.Mixed_7a = ReductionB(768)
self.Mixed_7b = InceptionC(1280)
self.Mixed_7c = InceptionC(2048)
self.fc = nn.Linear(2048, num_classes)
def forward(self, x):
if self.transform_input: # 1
x = x.clone()
x[:, 0] = x[:, 0] * (0.229 / 0.5) + (0.485 - 0.5) / 0.5
x[:, 1] = x[:, 1] * (0.224 / 0.5) + (0.456 - 0.5) / 0.5
x[:, 2] = x[:, 2] * (0.225 / 0.5) + (0.406 - 0.5) / 0.5
# 299 x 299 x 3
x = self.Conv2d_1a_3x3(x)
# 149 x 149 x 32
x = self.Conv2d_2a_3x3(x)
# 147 x 147 x 32
x = self.Conv2d_2b_3x3(x)
# 147 x 147 x 64
x = F.max_pool2d(x, kernel_size=3, stride=2)
# 73 x 73 x 64
x = self.Conv2d_3b_1x1(x)
# 73 x 73 x 80
x = self.Conv2d_4a_3x3(x)
# 71 x 71 x 192
x = F.max_pool2d(x, kernel_size=3, stride=2)
# 35 x 35 x 192
x = self.Mixed_5b(x)
# 35 x 35 x 256
x = self.Mixed_5c(x)
# 35 x 35 x 288
x = self.Mixed_5d(x)
# 35 x 35 x 288
x = self.Mixed_6a(x)
# 17 x 17 x 768
x = self.Mixed_6b(x)
# 17 x 17 x 768
x = self.Mixed_6c(x)
# 17 x 17 x 768
x = self.Mixed_6d(x)
# 17 x 17 x 768
x = self.Mixed_6e(x)
# 17 x 17 x 768
if self.training and self.aux_logits:
aux = self.AuxLogits(x)
# 17 x 17 x 768
x = self.Mixed_7a(x)
# 8 x 8 x 1280
x = self.Mixed_7b(x)
# 8 x 8 x 2048
x = self.Mixed_7c(x)
# 8 x 8 x 2048
x = F.avg_pool2d(x, kernel_size=5)
# 1 x 1 x 2048
x = F.dropout(x, training=self.training)
# 1 x 1 x 2048
x = x.view(x.size(0), -1)
# 2048
x = self.fc(x)
# 1000 (num_classes)
if self.training and self.aux_logits:
return x, aux
return x
# 统计模型参数量以及其他指标
import torchsummary
# 调用并将模型转移到GPU中
model = InceptionV3().to(device)
# 显示网络结构
torchsummary.summary(model, (3, 299, 299))
print(model)
def train(dataloader, model, loss_fn, optimizer):
size = len(dataloader.dataset) # 训练集的大小
num_batches = len(dataloader) # 批次数目, (size/batch_size,向上取整)
train_loss, train_acc = 0, 0 # 初始化训练损失和正确率
for X, y in dataloader: # 获取图片及其标签
X, y = X.to(device), y.to(device)
# 计算预测误差
pred = model(X) # 网络输出
loss = loss_fn(pred, y) # 计算网络输出和真实值之间的差距,targets为真实值,计算二者差值即为损失
# 反向传播
optimizer.zero_grad() # grad属性归零
loss.backward() # 反向传播
optimizer.step() # 每一步自动更新
# 记录acc与loss
train_acc += (pred.argmax(1) == y).type(torch.float).sum().item()
train_loss += loss.item()
train_acc /= size
train_loss /= num_batches
return train_acc, train_loss
def test(dataloader, model, loss_fn):
size = len(dataloader.dataset) # 测试集的大小
num_batches = len(dataloader) # 批次数目
test_loss, test_acc = 0, 0
# 当不进行训练时,停止梯度更新,节省计算内存消耗
with torch.no_grad():
for imgs, target in dataloader:
imgs, target = imgs.to(device), target.to(device)
# 计算loss
target_pred = model(imgs)
loss = loss_fn(target_pred, target)
test_loss += loss.item()
test_acc += (target_pred.argmax(1) == target).type(torch.float).sum().item()
test_acc /= size
test_loss /= num_batches
return test_acc, test_loss
import copy
optimizer = torch.optim.Adam(model.parameters(), lr=1e-4)
loss_fn = nn.CrossEntropyLoss() # 创建损失函数
epochs = 10
train_loss = []
train_acc = []
test_loss = []
test_acc = []
best_acc = 0 # 设置一个最佳准确率,作为最佳模型的判别指标
for epoch in range(epochs):
# 更新学习率(使用自定义学习率时使用)
# adjust_learning_rate(optimizer, epoch, learn_rate)
model.train()
epoch_train_acc, epoch_train_loss = train(train_dl, model, loss_fn, optimizer)
# scheduler.step() # 更新学习率(调用官方动态学习率接口时使用)
model.eval()
epoch_test_acc, epoch_test_loss = test(test_dl, model, loss_fn)
# 保存最佳模型到 best_model
if epoch_test_acc > best_acc:
best_acc = epoch_test_acc
best_model = copy.deepcopy(model)
train_acc.append(epoch_train_acc)
train_loss.append(epoch_train_loss)
test_acc.append(epoch_test_acc)
test_loss.append(epoch_test_loss)
# 获取当前的学习率
lr = optimizer.state_dict()['param_groups'][0]['lr']
template = ('Epoch:{:2d}, Train_acc:{:.1f}%, Train_loss:{:.3f}, Test_acc:{:.1f}%, Test_loss:{:.3f}, Lr:{:.2E}')
print(template.format(epoch + 1, epoch_train_acc * 100, epoch_train_loss,
epoch_test_acc * 100, epoch_test_loss, lr))
# 保存最佳模型到文件中
PATH = 'C:/Users/dell/PycharmProjects/pythonProject5/best_model.pth' # 保存的参数文件名
torch.save(model.state_dict(), PATH)
print('Done')
import matplotlib.pyplot as plt
# 隐藏警告
import warnings
warnings.filterwarnings("ignore") # 忽略警告信息
plt.rcParams['font.sans-serif'] = ['SimHei'] # 用来正常显示中文标签
plt.rcParams['axes.unicode_minus'] = False # 用来正常显示负号
plt.rcParams['figure.dpi'] = 100 # 分辨率
epochs_range = range(epochs)
plt.figure(figsize=(12, 3))
plt.subplot(1, 2, 1)
plt.plot(epochs_range, train_acc, label='Training Accuracy')
plt.plot(epochs_range, test_acc, label='Test Accuracy')
plt.legend(loc='lower right')
plt.title('Training and Validation Accuracy')
plt.subplot(1, 2, 2)
plt.plot(epochs_range, train_loss, label='Training Loss')
plt.plot(epochs_range, test_loss, label='Test Loss')
plt.legend(loc='upper right')
plt.title('Training and Validation Loss')
plt.show()
运行结果:
C:\Users\dell\anaconda3\envs\pytorch-gpu\python.exe C:\Users\dell\PycharmProjects\pythonProject5\models.py
Dataset ImageFolder
Number of datapoints: 1125
Root location: D:/研究生课题/深度学习-代码/3-date/weather_photos
StandardTransform
Transform: Compose(
Resize(size=[224, 224], interpolation=PIL.Image.BILINEAR)
ToTensor()
Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
)
<torch.utils.data.dataset.Subset object at 0x000002A91B643A08>
<torch.utils.data.dataset.Subset object at 0x000002A91E8B8848>
900 225
Shape of X [N, C, H, W]: torch.Size([32, 3, 224, 224])
Shape of y: torch.Size([32]) torch.int64
----------------------------------------------------------------
Layer (type) Output Shape Param #
================================================================
Conv2d-1 [-1, 32, 149, 149] 864
BatchNorm2d-2 [-1, 32, 149, 149] 64
ReLU-3 [-1, 32, 149, 149] 0
BasicConv2d-4 [-1, 32, 149, 149] 0
Conv2d-5 [-1, 32, 147, 147] 9,216
BatchNorm2d-6 [-1, 32, 147, 147] 64
ReLU-7 [-1, 32, 147, 147] 0
BasicConv2d-8 [-1, 32, 147, 147] 0
Conv2d-9 [-1, 64, 147, 147] 18,432
BatchNorm2d-10 [-1, 64, 147, 147] 128
ReLU-11 [-1, 64, 147, 147] 0
BasicConv2d-12 [-1, 64, 147, 147] 0
Conv2d-13 [-1, 80, 73, 73] 5,120
BatchNorm2d-14 [-1, 80, 73, 73] 160
ReLU-15 [-1, 80, 73, 73] 0
BasicConv2d-16 [-1, 80, 73, 73] 0
Conv2d-17 [-1, 192, 71, 71] 138,240
BatchNorm2d-18 [-1, 192, 71, 71] 384
ReLU-19 [-1, 192, 71, 71] 0
BasicConv2d-20 [-1, 192, 71, 71] 0
Conv2d-21 [-1, 64, 35, 35] 12,288
BatchNorm2d-22 [-1, 64, 35, 35] 128
ReLU-23 [-1, 64, 35, 35] 0
BasicConv2d-24 [-1, 64, 35, 35] 0
Conv2d-25 [-1, 48, 35, 35] 9,216
BatchNorm2d-26 [-1, 48, 35, 35] 96
ReLU-27 [-1, 48, 35, 35] 0
BasicConv2d-28 [-1, 48, 35, 35] 0
Conv2d-29 [-1, 64, 35, 35] 76,800
BatchNorm2d-30 [-1, 64, 35, 35] 128
ReLU-31 [-1, 64, 35, 35] 0
BasicConv2d-32 [-1, 64, 35, 35] 0
Conv2d-33 [-1, 64, 35, 35] 12,288
BatchNorm2d-34 [-1, 64, 35, 35] 128
ReLU-35 [-1, 64, 35, 35] 0
BasicConv2d-36 [-1, 64, 35, 35] 0
Conv2d-37 [-1, 96, 35, 35] 55,296
BatchNorm2d-38 [-1, 96, 35, 35] 192
ReLU-39 [-1, 96, 35, 35] 0
BasicConv2d-40 [-1, 96, 35, 35] 0
Conv2d-41 [-1, 96, 35, 35] 82,944
BatchNorm2d-42 [-1, 96, 35, 35] 192
ReLU-43 [-1, 96, 35, 35] 0
BasicConv2d-44 [-1, 96, 35, 35] 0
Conv2d-45 [-1, 32, 35, 35] 6,144
BatchNorm2d-46 [-1, 32, 35, 35] 64
ReLU-47 [-1, 32, 35, 35] 0
BasicConv2d-48 [-1, 32, 35, 35] 0
InceptionA-49 [-1, 256, 35, 35] 0
Conv2d-50 [-1, 64, 35, 35] 16,384
BatchNorm2d-51 [-1, 64, 35, 35] 128
ReLU-52 [-1, 64, 35, 35] 0
BasicConv2d-53 [-1, 64, 35, 35] 0
Conv2d-54 [-1, 48, 35, 35] 12,288
BatchNorm2d-55 [-1, 48, 35, 35] 96
ReLU-56 [-1, 48, 35, 35] 0
BasicConv2d-57 [-1, 48, 35, 35] 0
Conv2d-58 [-1, 64, 35, 35] 76,800
BatchNorm2d-59 [-1, 64, 35, 35] 128
ReLU-60 [-1, 64, 35, 35] 0
BasicConv2d-61 [-1, 64, 35, 35] 0
Conv2d-62 [-1, 64, 35, 35] 16,384
BatchNorm2d-63 [-1, 64, 35, 35] 128
ReLU-64 [-1, 64, 35, 35] 0
BasicConv2d-65 [-1, 64, 35, 35] 0
Conv2d-66 [-1, 96, 35, 35] 55,296
BatchNorm2d-67 [-1, 96, 35, 35] 192
ReLU-68 [-1, 96, 35, 35] 0
BasicConv2d-69 [-1, 96, 35, 35] 0
Conv2d-70 [-1, 96, 35, 35] 82,944
BatchNorm2d-71 [-1, 96, 35, 35] 192
ReLU-72 [-1, 96, 35, 35] 0
BasicConv2d-73 [-1, 96, 35, 35] 0
Conv2d-74 [-1, 64, 35, 35] 16,384
BatchNorm2d-75 [-1, 64, 35, 35] 128
ReLU-76 [-1, 64, 35, 35] 0
BasicConv2d-77 [-1, 64, 35, 35] 0
InceptionA-78 [-1, 288, 35, 35] 0
Conv2d-79 [-1, 64, 35, 35] 18,432
BatchNorm2d-80 [-1, 64, 35, 35] 128
ReLU-81 [-1, 64, 35, 35] 0
BasicConv2d-82 [-1, 64, 35, 35] 0
Conv2d-83 [-1, 48, 35, 35] 13,824
BatchNorm2d-84 [-1, 48, 35, 35] 96
ReLU-85 [-1, 48, 35, 35] 0
BasicConv2d-86 [-1, 48, 35, 35] 0
Conv2d-87 [-1, 64, 35, 35] 76,800
BatchNorm2d-88 [-1, 64, 35, 35] 128
ReLU-89 [-1, 64, 35, 35] 0
BasicConv2d-90 [-1, 64, 35, 35] 0
Conv2d-91 [-1, 64, 35, 35] 18,432
BatchNorm2d-92 [-1, 64, 35, 35] 128
ReLU-93 [-1, 64, 35, 35] 0
BasicConv2d-94 [-1, 64, 35, 35] 0
Conv2d-95 [-1, 96, 35, 35] 55,296
BatchNorm2d-96 [-1, 96, 35, 35] 192
ReLU-97 [-1, 96, 35, 35] 0
BasicConv2d-98 [-1, 96, 35, 35] 0
Conv2d-99 [-1, 96, 35, 35] 82,944
BatchNorm2d-100 [-1, 96, 35, 35] 192
ReLU-101 [-1, 96, 35, 35] 0
BasicConv2d-102 [-1, 96, 35, 35] 0
Conv2d-103 [-1, 64, 35, 35] 18,432
BatchNorm2d-104 [-1, 64, 35, 35] 128
ReLU-105 [-1, 64, 35, 35] 0
BasicConv2d-106 [-1, 64, 35, 35] 0
InceptionA-107 [-1, 288, 35, 35] 0
Conv2d-108 [-1, 384, 17, 17] 995,328
BatchNorm2d-109 [-1, 384, 17, 17] 768
ReLU-110 [-1, 384, 17, 17] 0
BasicConv2d-111 [-1, 384, 17, 17] 0
Conv2d-112 [-1, 64, 35, 35] 18,432
BatchNorm2d-113 [-1, 64, 35, 35] 128
ReLU-114 [-1, 64, 35, 35] 0
BasicConv2d-115 [-1, 64, 35, 35] 0
Conv2d-116 [-1, 96, 35, 35] 55,296
BatchNorm2d-117 [-1, 96, 35, 35] 192
ReLU-118 [-1, 96, 35, 35] 0
BasicConv2d-119 [-1, 96, 35, 35] 0
Conv2d-120 [-1, 96, 17, 17] 82,944
BatchNorm2d-121 [-1, 96, 17, 17] 192
ReLU-122 [-1, 96, 17, 17] 0
BasicConv2d-123 [-1, 96, 17, 17] 0
ReductionA-124 [-1, 768, 17, 17] 0
Conv2d-125 [-1, 192, 17, 17] 147,456
BatchNorm2d-126 [-1, 192, 17, 17] 384
ReLU-127 [-1, 192, 17, 17] 0
BasicConv2d-128 [-1, 192, 17, 17] 0
Conv2d-129 [-1, 128, 17, 17] 98,304
BatchNorm2d-130 [-1, 128, 17, 17] 256
ReLU-131 [-1, 128, 17, 17] 0
BasicConv2d-132 [-1, 128, 17, 17] 0
Conv2d-133 [-1, 128, 17, 17] 114,688
BatchNorm2d-134 [-1, 128, 17, 17] 256
ReLU-135 [-1, 128, 17, 17] 0
BasicConv2d-136 [-1, 128, 17, 17] 0
Conv2d-137 [-1, 192, 17, 17] 172,032
BatchNorm2d-138 [-1, 192, 17, 17] 384
ReLU-139 [-1, 192, 17, 17] 0
BasicConv2d-140 [-1, 192, 17, 17] 0
Conv2d-141 [-1, 128, 17, 17] 98,304
BatchNorm2d-142 [-1, 128, 17, 17] 256
ReLU-143 [-1, 128, 17, 17] 0
BasicConv2d-144 [-1, 128, 17, 17] 0
Conv2d-145 [-1, 128, 17, 17] 114,688
BatchNorm2d-146 [-1, 128, 17, 17] 256
ReLU-147 [-1, 128, 17, 17] 0
BasicConv2d-148 [-1, 128, 17, 17] 0
Conv2d-149 [-1, 128, 17, 17] 114,688
BatchNorm2d-150 [-1, 128, 17, 17] 256
ReLU-151 [-1, 128, 17, 17] 0
BasicConv2d-152 [-1, 128, 17, 17] 0
Conv2d-153 [-1, 128, 17, 17] 114,688
BatchNorm2d-154 [-1, 128, 17, 17] 256
ReLU-155 [-1, 128, 17, 17] 0
BasicConv2d-156 [-1, 128, 17, 17] 0
Conv2d-157 [-1, 192, 17, 17] 172,032
BatchNorm2d-158 [-1, 192, 17, 17] 384
ReLU-159 [-1, 192, 17, 17] 0
BasicConv2d-160 [-1, 192, 17, 17] 0
Conv2d-161 [-1, 192, 17, 17] 147,456
BatchNorm2d-162 [-1, 192, 17, 17] 384
ReLU-163 [-1, 192, 17, 17] 0
BasicConv2d-164 [-1, 192, 17, 17] 0
InceptionB-165 [-1, 768, 17, 17] 0
Conv2d-166 [-1, 192, 17, 17] 147,456
BatchNorm2d-167 [-1, 192, 17, 17] 384
ReLU-168 [-1, 192, 17, 17] 0
BasicConv2d-169 [-1, 192, 17, 17] 0
Conv2d-170 [-1, 160, 17, 17] 122,880
BatchNorm2d-171 [-1, 160, 17, 17] 320
ReLU-172 [-1, 160, 17, 17] 0
BasicConv2d-173 [-1, 160, 17, 17] 0
Conv2d-174 [-1, 160, 17, 17] 179,200
BatchNorm2d-175 [-1, 160, 17, 17] 320
ReLU-176 [-1, 160, 17, 17] 0
BasicConv2d-177 [-1, 160, 17, 17] 0
Conv2d-178 [-1, 192, 17, 17] 215,040
BatchNorm2d-179 [-1, 192, 17, 17] 384
ReLU-180 [-1, 192, 17, 17] 0
BasicConv2d-181 [-1, 192, 17, 17] 0
Conv2d-182 [-1, 160, 17, 17] 122,880
BatchNorm2d-183 [-1, 160, 17, 17] 320
ReLU-184 [-1, 160, 17, 17] 0
BasicConv2d-185 [-1, 160, 17, 17] 0
Conv2d-186 [-1, 160, 17, 17] 179,200
BatchNorm2d-187 [-1, 160, 17, 17] 320
ReLU-188 [-1, 160, 17, 17] 0
BasicConv2d-189 [-1, 160, 17, 17] 0
Conv2d-190 [-1, 160, 17, 17] 179,200
BatchNorm2d-191 [-1, 160, 17, 17] 320
ReLU-192 [-1, 160, 17, 17] 0
BasicConv2d-193 [-1, 160, 17, 17] 0
Conv2d-194 [-1, 160, 17, 17] 179,200
BatchNorm2d-195 [-1, 160, 17, 17] 320
ReLU-196 [-1, 160, 17, 17] 0
BasicConv2d-197 [-1, 160, 17, 17] 0
Conv2d-198 [-1, 192, 17, 17] 215,040
BatchNorm2d-199 [-1, 192, 17, 17] 384
ReLU-200 [-1, 192, 17, 17] 0
BasicConv2d-201 [-1, 192, 17, 17] 0
Conv2d-202 [-1, 192, 17, 17] 147,456
BatchNorm2d-203 [-1, 192, 17, 17] 384
ReLU-204 [-1, 192, 17, 17] 0
BasicConv2d-205 [-1, 192, 17, 17] 0
InceptionB-206 [-1, 768, 17, 17] 0
Conv2d-207 [-1, 192, 17, 17] 147,456
BatchNorm2d-208 [-1, 192, 17, 17] 384
ReLU-209 [-1, 192, 17, 17] 0
BasicConv2d-210 [-1, 192, 17, 17] 0
Conv2d-211 [-1, 160, 17, 17] 122,880
BatchNorm2d-212 [-1, 160, 17, 17] 320
ReLU-213 [-1, 160, 17, 17] 0
BasicConv2d-214 [-1, 160, 17, 17] 0
Conv2d-215 [-1, 160, 17, 17] 179,200
BatchNorm2d-216 [-1, 160, 17, 17] 320
ReLU-217 [-1, 160, 17, 17] 0
BasicConv2d-218 [-1, 160, 17, 17] 0
Conv2d-219 [-1, 192, 17, 17] 215,040
BatchNorm2d-220 [-1, 192, 17, 17] 384
ReLU-221 [-1, 192, 17, 17] 0
BasicConv2d-222 [-1, 192, 17, 17] 0
Conv2d-223 [-1, 160, 17, 17] 122,880
BatchNorm2d-224 [-1, 160, 17, 17] 320
ReLU-225 [-1, 160, 17, 17] 0
BasicConv2d-226 [-1, 160, 17, 17] 0
Conv2d-227 [-1, 160, 17, 17] 179,200
BatchNorm2d-228 [-1, 160, 17, 17] 320
ReLU-229 [-1, 160, 17, 17] 0
BasicConv2d-230 [-1, 160, 17, 17] 0
Conv2d-231 [-1, 160, 17, 17] 179,200
BatchNorm2d-232 [-1, 160, 17, 17] 320
ReLU-233 [-1, 160, 17, 17] 0
BasicConv2d-234 [-1, 160, 17, 17] 0
Conv2d-235 [-1, 160, 17, 17] 179,200
BatchNorm2d-236 [-1, 160, 17, 17] 320
ReLU-237 [-1, 160, 17, 17] 0
BasicConv2d-238 [-1, 160, 17, 17] 0
Conv2d-239 [-1, 192, 17, 17] 215,040
BatchNorm2d-240 [-1, 192, 17, 17] 384
ReLU-241 [-1, 192, 17, 17] 0
BasicConv2d-242 [-1, 192, 17, 17] 0
Conv2d-243 [-1, 192, 17, 17] 147,456
BatchNorm2d-244 [-1, 192, 17, 17] 384
ReLU-245 [-1, 192, 17, 17] 0
BasicConv2d-246 [-1, 192, 17, 17] 0
InceptionB-247 [-1, 768, 17, 17] 0
Conv2d-248 [-1, 192, 17, 17] 147,456
BatchNorm2d-249 [-1, 192, 17, 17] 384
ReLU-250 [-1, 192, 17, 17] 0
BasicConv2d-251 [-1, 192, 17, 17] 0
Conv2d-252 [-1, 192, 17, 17] 147,456
BatchNorm2d-253 [-1, 192, 17, 17] 384
ReLU-254 [-1, 192, 17, 17] 0
BasicConv2d-255 [-1, 192, 17, 17] 0
Conv2d-256 [-1, 192, 17, 17] 258,048
BatchNorm2d-257 [-1, 192, 17, 17] 384
ReLU-258 [-1, 192, 17, 17] 0
BasicConv2d-259 [-1, 192, 17, 17] 0
Conv2d-260 [-1, 192, 17, 17] 258,048
BatchNorm2d-261 [-1, 192, 17, 17] 384
ReLU-262 [-1, 192, 17, 17] 0
BasicConv2d-263 [-1, 192, 17, 17] 0
Conv2d-264 [-1, 192, 17, 17] 147,456
BatchNorm2d-265 [-1, 192, 17, 17] 384
ReLU-266 [-1, 192, 17, 17] 0
BasicConv2d-267 [-1, 192, 17, 17] 0
Conv2d-268 [-1, 192, 17, 17] 258,048
BatchNorm2d-269 [-1, 192, 17, 17] 384
ReLU-270 [-1, 192, 17, 17] 0
BasicConv2d-271 [-1, 192, 17, 17] 0
Conv2d-272 [-1, 192, 17, 17] 258,048
BatchNorm2d-273 [-1, 192, 17, 17] 384
ReLU-274 [-1, 192, 17, 17] 0
BasicConv2d-275 [-1, 192, 17, 17] 0
Conv2d-276 [-1, 192, 17, 17] 258,048
BatchNorm2d-277 [-1, 192, 17, 17] 384
ReLU-278 [-1, 192, 17, 17] 0
BasicConv2d-279 [-1, 192, 17, 17] 0
Conv2d-280 [-1, 192, 17, 17] 258,048
BatchNorm2d-281 [-1, 192, 17, 17] 384
ReLU-282 [-1, 192, 17, 17] 0
BasicConv2d-283 [-1, 192, 17, 17] 0
Conv2d-284 [-1, 192, 17, 17] 147,456
BatchNorm2d-285 [-1, 192, 17, 17] 384
ReLU-286 [-1, 192, 17, 17] 0
BasicConv2d-287 [-1, 192, 17, 17] 0
InceptionB-288 [-1, 768, 17, 17] 0
Conv2d-289 [-1, 192, 17, 17] 147,456
BatchNorm2d-290 [-1, 192, 17, 17] 384
ReLU-291 [-1, 192, 17, 17] 0
BasicConv2d-292 [-1, 192, 17, 17] 0
Conv2d-293 [-1, 320, 8, 8] 552,960
BatchNorm2d-294 [-1, 320, 8, 8] 640
ReLU-295 [-1, 320, 8, 8] 0
BasicConv2d-296 [-1, 320, 8, 8] 0
Conv2d-297 [-1, 192, 17, 17] 147,456
BatchNorm2d-298 [-1, 192, 17, 17] 384
ReLU-299 [-1, 192, 17, 17] 0
BasicConv2d-300 [-1, 192, 17, 17] 0
Conv2d-301 [-1, 192, 17, 17] 258,048
BatchNorm2d-302 [-1, 192, 17, 17] 384
ReLU-303 [-1, 192, 17, 17] 0
BasicConv2d-304 [-1, 192, 17, 17] 0
Conv2d-305 [-1, 192, 17, 17] 258,048
BatchNorm2d-306 [-1, 192, 17, 17] 384
ReLU-307 [-1, 192, 17, 17] 0
BasicConv2d-308 [-1, 192, 17, 17] 0
Conv2d-309 [-1, 192, 8, 8] 331,776
BatchNorm2d-310 [-1, 192, 8, 8] 384
ReLU-311 [-1, 192, 8, 8] 0
BasicConv2d-312 [-1, 192, 8, 8] 0
ReductionB-313 [-1, 1280, 8, 8] 0
Conv2d-314 [-1, 320, 8, 8] 409,600
BatchNorm2d-315 [-1, 320, 8, 8] 640
ReLU-316 [-1, 320, 8, 8] 0
BasicConv2d-317 [-1, 320, 8, 8] 0
Conv2d-318 [-1, 384, 8, 8] 491,520
BatchNorm2d-319 [-1, 384, 8, 8] 768
ReLU-320 [-1, 384, 8, 8] 0
BasicConv2d-321 [-1, 384, 8, 8] 0
Conv2d-322 [-1, 384, 8, 8] 442,368
BatchNorm2d-323 [-1, 384, 8, 8] 768
ReLU-324 [-1, 384, 8, 8] 0
BasicConv2d-325 [-1, 384, 8, 8] 0
Conv2d-326 [-1, 384, 8, 8] 442,368
BatchNorm2d-327 [-1, 384, 8, 8] 768
ReLU-328 [-1, 384, 8, 8] 0
BasicConv2d-329 [-1, 384, 8, 8] 0
Conv2d-330 [-1, 448, 8, 8] 573,440
BatchNorm2d-331 [-1, 448, 8, 8] 896
ReLU-332 [-1, 448, 8, 8] 0
BasicConv2d-333 [-1, 448, 8, 8] 0
Conv2d-334 [-1, 384, 8, 8] 1,548,288
BatchNorm2d-335 [-1, 384, 8, 8] 768
ReLU-336 [-1, 384, 8, 8] 0
BasicConv2d-337 [-1, 384, 8, 8] 0
Conv2d-338 [-1, 384, 8, 8] 442,368
BatchNorm2d-339 [-1, 384, 8, 8] 768
ReLU-340 [-1, 384, 8, 8] 0
BasicConv2d-341 [-1, 384, 8, 8] 0
Conv2d-342 [-1, 384, 8, 8] 442,368
BatchNorm2d-343 [-1, 384, 8, 8] 768
ReLU-344 [-1, 384, 8, 8] 0
BasicConv2d-345 [-1, 384, 8, 8] 0
Conv2d-346 [-1, 192, 8, 8] 245,760
BatchNorm2d-347 [-1, 192, 8, 8] 384
ReLU-348 [-1, 192, 8, 8] 0
BasicConv2d-349 [-1, 192, 8, 8] 0
InceptionC-350 [-1, 2048, 8, 8] 0
Conv2d-351 [-1, 320, 8, 8] 655,360
BatchNorm2d-352 [-1, 320, 8, 8] 640
ReLU-353 [-1, 320, 8, 8] 0
BasicConv2d-354 [-1, 320, 8, 8] 0
Conv2d-355 [-1, 384, 8, 8] 786,432
BatchNorm2d-356 [-1, 384, 8, 8] 768
ReLU-357 [-1, 384, 8, 8] 0
BasicConv2d-358 [-1, 384, 8, 8] 0
Conv2d-359 [-1, 384, 8, 8] 442,368
BatchNorm2d-360 [-1, 384, 8, 8] 768
ReLU-361 [-1, 384, 8, 8] 0
BasicConv2d-362 [-1, 384, 8, 8] 0
Conv2d-363 [-1, 384, 8, 8] 442,368
BatchNorm2d-364 [-1, 384, 8, 8] 768
ReLU-365 [-1, 384, 8, 8] 0
BasicConv2d-366 [-1, 384, 8, 8] 0
Conv2d-367 [-1, 448, 8, 8] 917,504
BatchNorm2d-368 [-1, 448, 8, 8] 896
ReLU-369 [-1, 448, 8, 8] 0
BasicConv2d-370 [-1, 448, 8, 8] 0
Conv2d-371 [-1, 384, 8, 8] 1,548,288
BatchNorm2d-372 [-1, 384, 8, 8] 768
ReLU-373 [-1, 384, 8, 8] 0
BasicConv2d-374 [-1, 384, 8, 8] 0
Conv2d-375 [-1, 384, 8, 8] 442,368
BatchNorm2d-376 [-1, 384, 8, 8] 768
ReLU-377 [-1, 384, 8, 8] 0
BasicConv2d-378 [-1, 384, 8, 8] 0
Conv2d-379 [-1, 384, 8, 8] 442,368
BatchNorm2d-380 [-1, 384, 8, 8] 768
ReLU-381 [-1, 384, 8, 8] 0
BasicConv2d-382 [-1, 384, 8, 8] 0
Conv2d-383 [-1, 192, 8, 8] 393,216
BatchNorm2d-384 [-1, 192, 8, 8] 384
ReLU-385 [-1, 192, 8, 8] 0
BasicConv2d-386 [-1, 192, 8, 8] 0
InceptionC-387 [-1, 2048, 8, 8] 0
Linear-388 [-1, 1000] 2,049,000
================================================================
Total params: 23,834,568
Trainable params: 23,834,568
Non-trainable params: 0
----------------------------------------------------------------
Input size (MB): 1.02
Forward/backward pass size (MB): 292.54
Params size (MB): 90.92
Estimated Total Size (MB): 384.49
----------------------------------------------------------------
InceptionV3(
(Conv2d_1a_3x3): BasicConv2d(
(conv): Conv2d(3, 32, kernel_size=(3, 3), stride=(2, 2), bias=False)
(norm): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(Conv2d_2a_3x3): BasicConv2d(
(conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), bias=False)
(norm): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(Conv2d_2b_3x3): BasicConv2d(
(conv): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(norm): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(Conv2d_3b_1x1): BasicConv2d(
(conv): Conv2d(64, 80, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(80, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(Conv2d_4a_3x3): BasicConv2d(
(conv): Conv2d(80, 192, kernel_size=(3, 3), stride=(1, 1), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(Mixed_5b): InceptionA(
(branch1x1): BasicConv2d(
(conv): Conv2d(192, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch5x5_1): BasicConv2d(
(conv): Conv2d(192, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(48, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch5x5_2): BasicConv2d(
(conv): Conv2d(48, 64, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), bias=False)
(norm): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_1): BasicConv2d(
(conv): Conv2d(192, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_2): BasicConv2d(
(conv): Conv2d(64, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(norm): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_3): BasicConv2d(
(conv): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(norm): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch_pool): BasicConv2d(
(conv): Conv2d(192, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(Mixed_5c): InceptionA(
(branch1x1): BasicConv2d(
(conv): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch5x5_1): BasicConv2d(
(conv): Conv2d(256, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(48, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch5x5_2): BasicConv2d(
(conv): Conv2d(48, 64, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), bias=False)
(norm): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_1): BasicConv2d(
(conv): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_2): BasicConv2d(
(conv): Conv2d(64, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(norm): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_3): BasicConv2d(
(conv): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(norm): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch_pool): BasicConv2d(
(conv): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(Mixed_5d): InceptionA(
(branch1x1): BasicConv2d(
(conv): Conv2d(288, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch5x5_1): BasicConv2d(
(conv): Conv2d(288, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(48, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch5x5_2): BasicConv2d(
(conv): Conv2d(48, 64, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), bias=False)
(norm): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_1): BasicConv2d(
(conv): Conv2d(288, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_2): BasicConv2d(
(conv): Conv2d(64, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(norm): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_3): BasicConv2d(
(conv): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(norm): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch_pool): BasicConv2d(
(conv): Conv2d(288, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(Mixed_6a): ReductionA(
(branch3x3): BasicConv2d(
(conv): Conv2d(288, 384, kernel_size=(3, 3), stride=(2, 2), bias=False)
(norm): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_1): BasicConv2d(
(conv): Conv2d(288, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_2): BasicConv2d(
(conv): Conv2d(64, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(norm): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_3): BasicConv2d(
(conv): Conv2d(96, 96, kernel_size=(3, 3), stride=(2, 2), bias=False)
(norm): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(Mixed_6b): InceptionB(
(branch1x1): BasicConv2d(
(conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7_1): BasicConv2d(
(conv): Conv2d(768, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7_2): BasicConv2d(
(conv): Conv2d(128, 128, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)
(norm): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7_3): BasicConv2d(
(conv): Conv2d(128, 192, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_1): BasicConv2d(
(conv): Conv2d(768, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_2): BasicConv2d(
(conv): Conv2d(128, 128, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)
(norm): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_3): BasicConv2d(
(conv): Conv2d(128, 128, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)
(norm): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_4): BasicConv2d(
(conv): Conv2d(128, 128, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)
(norm): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_5): BasicConv2d(
(conv): Conv2d(128, 192, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch_pool): BasicConv2d(
(conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(Mixed_6c): InceptionB(
(branch1x1): BasicConv2d(
(conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7_1): BasicConv2d(
(conv): Conv2d(768, 160, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7_2): BasicConv2d(
(conv): Conv2d(160, 160, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)
(norm): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7_3): BasicConv2d(
(conv): Conv2d(160, 192, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_1): BasicConv2d(
(conv): Conv2d(768, 160, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_2): BasicConv2d(
(conv): Conv2d(160, 160, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)
(norm): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_3): BasicConv2d(
(conv): Conv2d(160, 160, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)
(norm): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_4): BasicConv2d(
(conv): Conv2d(160, 160, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)
(norm): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_5): BasicConv2d(
(conv): Conv2d(160, 192, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch_pool): BasicConv2d(
(conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(Mixed_6d): InceptionB(
(branch1x1): BasicConv2d(
(conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7_1): BasicConv2d(
(conv): Conv2d(768, 160, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7_2): BasicConv2d(
(conv): Conv2d(160, 160, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)
(norm): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7_3): BasicConv2d(
(conv): Conv2d(160, 192, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_1): BasicConv2d(
(conv): Conv2d(768, 160, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_2): BasicConv2d(
(conv): Conv2d(160, 160, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)
(norm): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_3): BasicConv2d(
(conv): Conv2d(160, 160, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)
(norm): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_4): BasicConv2d(
(conv): Conv2d(160, 160, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)
(norm): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_5): BasicConv2d(
(conv): Conv2d(160, 192, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch_pool): BasicConv2d(
(conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(Mixed_6e): InceptionB(
(branch1x1): BasicConv2d(
(conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7_1): BasicConv2d(
(conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7_2): BasicConv2d(
(conv): Conv2d(192, 192, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7_3): BasicConv2d(
(conv): Conv2d(192, 192, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_1): BasicConv2d(
(conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_2): BasicConv2d(
(conv): Conv2d(192, 192, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_3): BasicConv2d(
(conv): Conv2d(192, 192, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_4): BasicConv2d(
(conv): Conv2d(192, 192, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7dbl_5): BasicConv2d(
(conv): Conv2d(192, 192, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch_pool): BasicConv2d(
(conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(Mixed_7a): ReductionB(
(branch3x3_1): BasicConv2d(
(conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3_2): BasicConv2d(
(conv): Conv2d(192, 320, kernel_size=(3, 3), stride=(2, 2), bias=False)
(norm): BatchNorm2d(320, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7x3_1): BasicConv2d(
(conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7x3_2): BasicConv2d(
(conv): Conv2d(192, 192, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7x3_3): BasicConv2d(
(conv): Conv2d(192, 192, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch7x7x3_4): BasicConv2d(
(conv): Conv2d(192, 192, kernel_size=(3, 3), stride=(2, 2), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(Mixed_7b): InceptionC(
(branch1x1): BasicConv2d(
(conv): Conv2d(1280, 320, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(320, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3_1): BasicConv2d(
(conv): Conv2d(1280, 384, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3_2a): BasicConv2d(
(conv): Conv2d(384, 384, kernel_size=(1, 3), stride=(1, 1), padding=(0, 1), bias=False)
(norm): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3_2b): BasicConv2d(
(conv): Conv2d(384, 384, kernel_size=(3, 1), stride=(1, 1), padding=(1, 0), bias=False)
(norm): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_1): BasicConv2d(
(conv): Conv2d(1280, 448, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(448, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_2): BasicConv2d(
(conv): Conv2d(448, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(norm): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_3a): BasicConv2d(
(conv): Conv2d(384, 384, kernel_size=(1, 3), stride=(1, 1), padding=(0, 1), bias=False)
(norm): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_3b): BasicConv2d(
(conv): Conv2d(384, 384, kernel_size=(3, 1), stride=(1, 1), padding=(1, 0), bias=False)
(norm): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch_pool): BasicConv2d(
(conv): Conv2d(1280, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(Mixed_7c): InceptionC(
(branch1x1): BasicConv2d(
(conv): Conv2d(2048, 320, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(320, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3_1): BasicConv2d(
(conv): Conv2d(2048, 384, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3_2a): BasicConv2d(
(conv): Conv2d(384, 384, kernel_size=(1, 3), stride=(1, 1), padding=(0, 1), bias=False)
(norm): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3_2b): BasicConv2d(
(conv): Conv2d(384, 384, kernel_size=(3, 1), stride=(1, 1), padding=(1, 0), bias=False)
(norm): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_1): BasicConv2d(
(conv): Conv2d(2048, 448, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(448, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_2): BasicConv2d(
(conv): Conv2d(448, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(norm): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_3a): BasicConv2d(
(conv): Conv2d(384, 384, kernel_size=(1, 3), stride=(1, 1), padding=(0, 1), bias=False)
(norm): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch3x3dbl_3b): BasicConv2d(
(conv): Conv2d(384, 384, kernel_size=(3, 1), stride=(1, 1), padding=(1, 0), bias=False)
(norm): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(branch_pool): BasicConv2d(
(conv): Conv2d(2048, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
(norm): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(fc): Linear(in_features=2048, out_features=1000, bias=True)
)
Epoch: 1, Train_acc:68.8%, Train_loss:2.668, Test_acc:27.6%, Test_loss:5.936, Lr:1.00E-04
Epoch: 2, Train_acc:86.9%, Train_loss:0.558, Test_acc:88.4%, Test_loss:0.505, Lr:1.00E-04
Epoch: 3, Train_acc:88.4%, Train_loss:0.551, Test_acc:83.6%, Test_loss:0.630, Lr:1.00E-04
Epoch: 4, Train_acc:90.9%, Train_loss:0.358, Test_acc:91.1%, Test_loss:0.268, Lr:1.00E-04
Epoch: 5, Train_acc:90.8%, Train_loss:0.345, Test_acc:87.1%, Test_loss:0.409, Lr:1.00E-04
Epoch: 6, Train_acc:90.4%, Train_loss:0.275, Test_acc:93.3%, Test_loss:0.167, Lr:1.00E-04
Epoch: 7, Train_acc:94.6%, Train_loss:0.190, Test_acc:90.7%, Test_loss:0.252, Lr:1.00E-04
Epoch: 8, Train_acc:92.8%, Train_loss:0.215, Test_acc:92.0%, Test_loss:0.220, Lr:1.00E-04
Epoch: 9, Train_acc:94.9%, Train_loss:0.136, Test_acc:89.8%, Test_loss:0.339, Lr:1.00E-04
Epoch:10, Train_acc:95.2%, Train_loss:0.138, Test_acc:92.9%, Test_loss:0.181, Lr:1.00E-04
Done
进程已结束,退出代码为 0
运行结果截图:
总结:InceptionV3相对于InceptionV1的主要改进有:将大卷积核分解为多个小卷积核,减少计算量和参数量。增加了网络层数,增强了特征表示能力。提高训练速度和网络稳定性。通过平滑标签防止过拟合,提高泛化能力。在Inception模块中结合卷积和池化操作,更好地提取多尺度特征。这些改进提高了InceptionV3的计算效率和性能。