VGG16

修改默认路径

import os
os.environ['TORCH_HOME'] = r'D:\Pytorch\pythonProject\vgg16'  # 下载位置

太大了（140多G）不提供直接下载

train_set = torchvision.datasets.ImageNet(root='./data_image_net', split='train', download=True, transform=torchvision.transforms.ToTensor())

是否预训练

不预训练：采用随机参数
预训练：采用训练好的参数

第一次

第二次

vgg16_false = torchvision.models.vgg16(weights=None)
vgg16_true = torchvision.models.vgg16(weights='DEFAULT')  # or weights='IMAGENET1K_V1'

完整代码

import torchvision
import osos.environ['TORCH_HOME'] = r'D:\Pytorch\pythonProject\vgg16'  # 下载位置# train_set = torchvision.datasets.ImageNet(root='./data_image_net', split='train', download=True
#                                           , transform=torchvision.transforms.ToTensor())vgg16_false = torchvision.models.vgg16(weights=None)
vgg16_true = torchvision.models.vgg16(weights='DEFAULT')  # or weights='IMAGENET1K_V1'
print(vgg16_true)

加一层线性层-nn.Linear

vgg16_true.add_module('add_linear', nn.Linear(1000, 10))

如果想加到classifier里面

vgg16_true.classifier.add_module('add_linear', nn.Linear(1000, 10))

修改神经网络某层

vgg16_false.classifier[6] = nn.Linear(4096, 10)

改之前

print(vgg16_false)

改之后