0
0

10. 网络模型使用及修改

10.1 下载网络模型

  1. import torchvision
  3. #trauin_data = torchvision.datasets.ImageNet("./dataset",split="train",download=True,transform=torchvision.transforms.ToTensor())   # 这个数据集没有办法再公开的访问了    
  4. vgg16_true = torchvision.models.vgg16(pretrained=True) # 下载卷积层对应的参数是多少、池化层对应的参数时多少，这些参数时ImageNet训练好了的
  5. vgg16_false = torchvision.models.vgg16(pretrained=False) # 没有预训练的参数
  6. print("ok")
  7. print(vgg16_true)

结果：

  1. ok
  2. VGG(
  3.   (features): Sequential(
  4.     (0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  5.     (1): ReLU(inplace=True)
  6.     (2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  7.     (3): ReLU(inplace=True)
  8.     (4): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  9.     (5): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  10.     (6): ReLU(inplace=True)
  11.     (7): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  12.     (8): ReLU(inplace=True)
  13.     (9): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  14.     (10): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  15.     (11): ReLU(inplace=True)
  16.     (12): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  17.     (13): ReLU(inplace=True)
  18.     (14): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  19.     (15): ReLU(inplace=True)
  20.     (16): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  21.     (17): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  22.     (18): ReLU(inplace=True)
  23.     (19): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  24.     (20): ReLU(inplace=True)
  25.     (21): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  26.     (22): ReLU(inplace=True)
  27.     (23): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  28.     (24): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  29.     (25): ReLU(inplace=True)
  30.     (26): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  31.     (27): ReLU(inplace=True)
  32.     (28): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  33.     (29): ReLU(inplace=True)
  34.     (30): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  35.   )
  36.   (avgpool): AdaptiveAvgPool2d(output_size=(7, 7))
  37.   (classifier): Sequential(
  38.     (0): Linear(in_features=25088, out_features=4096, bias=True)
  39.     (1): ReLU(inplace=True)
  40.     (2): Dropout(p=0.5, inplace=False)
  41.     (3): Linear(in_features=4096, out_features=4096, bias=True)
  42.     (4): ReLU(inplace=True)
  43.     (5): Dropout(p=0.5, inplace=False)
  44.     (6): Linear(in_features=4096, out_features=1000, bias=True)
  45.   )

10.2 查看函数用法

  1. import torchvision
  2. help(torchvision.models.vgg16)

结果：

  1. Help on function vgg16 in module torchvision.models.vgg:
  3. vgg16(pretrained:bool=False, progress:bool=True, **kwargs:Any) -> torchvision.models.vgg.VGG
  4.     VGG 16-layer model (configuration "D")
  5.     `"Very Deep Convolutional Networks For Large-Scale Image Recognition" <https://arxiv.org/pdf/1409.1556.pdf>`_.
  6.     The required minimum input size of the model is 32x32.
  7.     
  8.     Args:
  9.         pretrained (bool): If True, returns a model pre-trained on ImageNet
  10.         progress (bool): If True, displays a progress bar of the download to stderr

10.3 网络模型添加

  1. import torchvision
  2. from torch import nn
  4. dataset = torchvision.datasets.CIFAR10("./dataset",train=True,transform=torchvision.transforms.ToTensor(),download=True)       
  5. vgg16_true = torchvision.models.vgg16(pretrained=True) # 下载卷积层对应的参数是多少、池化层对应的参数时多少，这些参数时ImageNet训练好了的
  6. vgg16_true.add_module('add_linear',nn.Linear(1000,10)) # 在VGG16后面添加一个线性层，使得输出为适应CIFAR10的输出，CIFAR10需要输出10个种类
  8. print(vgg16_true)

结果：

  1. Files already downloaded and verified
  2. VGG(
  3.   (features): Sequential(
  4.     (0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  5.     (1): ReLU(inplace=True)
  6.     (2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  7.     (3): ReLU(inplace=True)
  8.     (4): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  9.     (5): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  10.     (6): ReLU(inplace=True)
  11.     (7): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  12.     (8): ReLU(inplace=True)
  13.     (9): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  14.     (10): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  15.     (11): ReLU(inplace=True)
  16.     (12): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  17.     (13): ReLU(inplace=True)
  18.     (14): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  19.     (15): ReLU(inplace=True)
  20.     (16): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  21.     (17): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  22.     (18): ReLU(inplace=True)
  23.     (19): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  24.     (20): ReLU(inplace=True)
  25.     (21): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  26.     (22): ReLU(inplace=True)
  27.     (23): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  28.     (24): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  29.     (25): ReLU(inplace=True)
  30.     (26): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  31.     (27): ReLU(inplace=True)
  32.     (28): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  33.     (29): ReLU(inplace=True)
  34.     (30): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  35.   )
  36.   (avgpool): AdaptiveAvgPool2d(output_size=(7, 7))
  37.   (classifier): Sequential(
  38.     (0): Linear(in_features=25088, out_features=4096, bias=True)
  39.     (1): ReLU(inplace=True)
  40.     (2): Dropout(p=0.5, inplace=False)
  41.     (3): Linear(in_features=4096, out_features=4096, bias=True)
  42.     (4): ReLU(inplace=True)
  43.     (5): Dropout(p=0.5, inplace=False)
  44.     (6): Linear(in_features=4096, out_features=1000, bias=True)
  45.   )
  46.   (add_linear): Linear(in_features=1000, out_features=10, bias=True)
  47. )

10.4 网络模型修改

  1. import torchvision
  2. from torch import nn
  4. vgg16_false = torchvision.models.vgg16(pretrained=False) # 没有预训练的参数     
  5. print(vgg16_false)
  6. vgg16_false.classifier[6] = nn.Linear(4096,10)
  7. print(vgg16_false)

结果：

  1. VGG(
  2.   (features): Sequential(
  3.     (0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  4.     (1): ReLU(inplace=True)
  5.     (2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  6.     (3): ReLU(inplace=True)
  7.     (4): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  8.     (5): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  9.     (6): ReLU(inplace=True)
  10.     (7): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  11.     (8): ReLU(inplace=True)
  12.     (9): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  13.     (10): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  14.     (11): ReLU(inplace=True)
  15.     (12): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  16.     (13): ReLU(inplace=True)
  17.     (14): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  18.     (15): ReLU(inplace=True)
  19.     (16): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  20.     (17): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  21.     (18): ReLU(inplace=True)
  22.     (19): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  23.     (20): ReLU(inplace=True)
  24.     (21): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  25.     (22): ReLU(inplace=True)
  26.     (23): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  27.     (24): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  28.     (25): ReLU(inplace=True)
  29.     (26): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  30.     (27): ReLU(inplace=True)
  31.     (28): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  32.     (29): ReLU(inplace=True)
  33.     (30): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  34.   )
  35.   (avgpool): AdaptiveAvgPool2d(output_size=(7, 7))
  36.   (classifier): Sequential(
  37.     (0): Linear(in_features=25088, out_features=4096, bias=True)
  38.     (1): ReLU(inplace=True)
  39.     (2): Dropout(p=0.5, inplace=False)
  40.     (3): Linear(in_features=4096, out_features=4096, bias=True)
  41.     (4): ReLU(inplace=True)
  42.     (5): Dropout(p=0.5, inplace=False)
  43.     (6): Linear(in_features=4096, out_features=1000, bias=True)
  44.   )
  45. )
  46. VGG(
  47.   (features): Sequential(
  48.     (0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  49.     (1): ReLU(inplace=True)
  50.     (2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  51.     (3): ReLU(inplace=True)
  52.     (4): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  53.     (5): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  54.     (6): ReLU(inplace=True)
  55.     (7): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  56.     (8): ReLU(inplace=True)
  57.     (9): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  58.     (10): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  59.     (11): ReLU(inplace=True)
  60.     (12): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  61.     (13): ReLU(inplace=True)
  62.     (14): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  63.     (15): ReLU(inplace=True)
  64.     (16): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  65.     (17): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  66.     (18): ReLU(inplace=True)
  67.     (19): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  68.     (20): ReLU(inplace=True)
  69.     (21): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  70.     (22): ReLU(inplace=True)
  71.     (23): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  72.     (24): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  73.     (25): ReLU(inplace=True)
  74.     (26): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  75.     (27): ReLU(inplace=True)
  76.     (28): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  77.     (29): ReLU(inplace=True)
  78.     (30): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  79.   )
  80.   (avgpool): AdaptiveAvgPool2d(output_size=(7, 7))
  81.   (classifier): Sequential(
  82.     (0): Linear(in_features=25088, out_features=4096, bias=True)
  83.     (1): ReLU(inplace=True)
  84.     (2): Dropout(p=0.5, inplace=False)
  85.     (3): Linear(in_features=4096, out_features=4096, bias=True)
  86.     (4): ReLU(inplace=True)
  87.     (5): Dropout(p=0.5, inplace=False)
  88.     (6): Linear(in_features=4096, out_features=10, bias=True)
  89.   )
  90. )
标签： python 深度学习 人工智能
本文转载自: https://blog.csdn.net/qq_54932411/article/details/132513293
版权归原作者 Gosling123456 所有， 如有侵权，请联系我们删除。
登录后发布评论

“10. 网络模型使用及修改”的评论:

还没有评论
关于作者
...
overfit同步小助手
文章同步
相关阅读

网络安全法-网络运行安全

使用selenium/drissionpage时如何阻止chrome自动跳转http到https

docker desktop 里部署的Open WebUI 管理员密码忘记了的处理方法

在ubuntu20.04中搭建onsite比赛运行环境

利用开源的低代码表单设计器FcDesigner高效管理和渲染复杂表单结构

Kafka学习笔记

【前端】浏览器输入url到页面呈现发生了什么？