利用高斯核卷积对MINIST数据集进行去噪
import torch
import torchvision
from torch.autograd import Variable
from torchvision import datasets, transforms
from torch.utils.data import DataLoader
import cv2
from torch import nn
import numpy as np
import torch.nn.functional as F
import advertorch.defenses as defenses
from numpy import *
seed = 2014
torch.manual_seed(seed)
np.random.seed(seed) # Numpy module.
random.seed(seed) # Python random module.
torch.manual_seed(seed)
train_dataset = datasets.FashionMNIST('./fashionmnist_data/', train=True, download=True,
transform=transforms.Compose([
transforms.ToTensor(),
]))
train_loader = DataLoader(dataset = train_dataset, batch_size = 500, shuffle = True)
test_loader = torch.utils.data.DataLoader(
datasets.FashionMNIST('./fashionmnist_data/', train=False, transform=transforms.Compose([
transforms.ToTensor(),
])),batch_size=1, shuffle=True)
epoch = 12
class Linear_cliassifer(torch.nn.Module):
def __init__(self) :
super(Linear_cliassifer, self).__init__()
self.Gs = defenses.GaussianSmoothing2D(3, 1, 3)
self.Line1 = torch.nn.Linear(28 * 28, 10)
def forward(self, x):
x = self.Gs(x)
x = self.Line1(x.view(-1, 28 * 28))
return x
net = Linear_cliassifer()
cost = torch.nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(net.parameters(), lr=0.001)
for k in range(epoch):
sum_loss = 0.0
train_correct = 0
for i, data in enumerate(train_loader, 0):
inputs, labels = data
optimizer.zero_grad()
outputs = net(inputs)
loss = cost(outputs, labels)
loss.backward()
optimizer.step()
print(loss)
_, id = torch.max(outputs.data, 1)
sum_loss += loss.data
train_correct += torch.sum(id == labels.data)
#print('[%d,%d] loss:%.03f' % (k + 1, k, sum_loss / len(train_loader)))
print(' correct:%.03f%%' % (100 * train_correct / len(train_dataset)))
torch.save(net.state_dict(), 'model/fasion_BL.pt')
本文地址：https://blog.csdn.net/qq_23144435/article/details/107430946