Eye_Blinking_PYTORCH

import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torchvision import datasets, transforms
from torch.autograd import Variable
from torch.utils.data import Dataset
import os
import glob
import matplotlib.pyplot as plt
import numpy as np

imgs_open = glob.glob('datasets/openRightEyes/*.jpg')
imgs_closed = glob.glob('datasets/closedRightEyes/*.jpg')

batch_size = 50
iterations = 20

class Train_Dataset(Dataset):
    def __init__(self):
        self.images = []
        self.labels = []

        for i in range(0, 1000):
            image = plt.imread(imgs_open[i])
            image = np.reshape(image, [1, 24, 24])
            self.images.append(image)
            self.labels.append(0)

        for i in range(0, 1000):
            image = plt.imread(imgs_closed[i])
            image = np.reshape(image, [1, 24, 24])
            self.images.append(image)
            self.labels.append(1)

        self.images = np.array(self.images, dtype=np.float32)
        self.labels = np.array(self.labels, dtype=np.int64)

        self.images = torch.from_numpy(self.images)
        self.labels = torch.from_numpy(self.labels)

    def __getitem__(self, index):
        return self.images[index], self.labels[index]

    def __len__(self):
        return 2000


class Test_Dataset(Dataset):
    def __init__(self):
        self.images = []
        self.labels = []

        for i in range(1000, len(imgs_open)):
            image = plt.imread(imgs_open[i])
            image = np.reshape(image, [1, 24, 24])
            self.images.append(image)
            self.labels.append(0)

        for i in range(1000, len(imgs_closed)):
            image = plt.imread(imgs_closed[i])
            image = np.reshape(image, [1, 24, 24])
            self.images.append(image)
            self.labels.append(1)

        self.images = np.array(self.images, dtype=np.float32)
        self.labels = np.array(self.labels, dtype=np.int64)

        self.images = torch.from_numpy(self.images)
        self.labels = torch.from_numpy(self.labels)
    def __getitem__(self, index):
        return self.images[index], self.labels[index]

    def __len__(self):
        return 423

train_dataset = Train_Dataset()
test_dataset = Test_Dataset()

train_loader = torch.utils.data.DataLoader(dataset=train_dataset,
                                           batch_size=batch_size,
                                           shuffle=True)

test_loader = torch.utils.data.DataLoader(dataset=test_dataset,
                                          batch_size=batch_size,
                                          shuffle=False)







class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        # 输入1通道，输出32通道，kernel 3*3
        self.conv1 = nn.Conv2d(1, 32, kernel_size=3, padding=1)
        self.conv2 = nn.Conv2d(32, 50, kernel_size=3, padding=1)
        self.mp = nn.MaxPool2d(2)
        # fully connect
        self.fc1 = nn.Linear(1800, 1024)
        self.fc2 = nn.Linear(1024, 2)

    def forward(self, x):

        x = F.relu(self.conv1(x))
        x = self.mp(x)
        x = F.relu(self.conv2(x))
        x = self.mp(x)
        b_n = x.size(0)
        x = x.view(b_n, -1)
        x = F.relu(self.fc1(x))
        x = self.fc2(x)
        return F.log_softmax(x)


net = Net()
optimizer = optim.SGD(net.parameters(), lr=0.001, momentum=0.9)


for i in range(iterations):

    #Train
    for batch_idx, (data, target) in enumerate(train_loader):
        data, target = Variable(data), Variable(target)
        optimizer.zero_grad()
        output = net(data)
        loss = F.nll_loss(output, target)
        loss.backward()
        optimizer.step()

        pred = output.data.max(1, keepdim=True)[1]
        accuracy_ = pred.eq(target.data.view_as(pred)).float()
        accuracy = accuracy_.mean()

        if batch_idx % 5 == 0:
            print('Train Epoch: %d  [%d/%d ]  %f \tTrainLoss: %f\tTrainAcc:%f'%(
                i, batch_idx * len(data), len(train_loader.dataset),
                       1.0 * batch_idx / len(train_loader), loss.data, accuracy))

    # Test
    accuracy_all = 0
    for batch_idx, (data, target) in enumerate(test_loader):
        data, target = Variable(data), Variable(target)
        output = net(data)
        pred = output.data.max(1, keepdim=True)[1]
        accuracy_ = pred.eq(target.data.view_as(pred))
        accuracy = accuracy_.sum()
        accuracy_all += accuracy
    x1 = len(test_loader.dataset)
    accuracy_mean = accuracy_all.float()/len(test_loader.dataset)

    print('Train Epoch: %d\tTestAcc: %f'%(
                i,  accuracy_mean))

数据集已经上传到了我的CSDN的资料库里