深度學習識別CIFAR10：pytorch訓練LeNet、AlexNet、VGG19實現及比較（一）

版權聲明：本文為博主原創文章，歡迎轉載，並請注明出處。聯系方式：[e-mail]

前面幾篇文章介紹了MINIST，對這種簡單圖片的識別，LeNet-5可以達到99%的識別率。

CIFAR10是另一個著名的深度學習圖像分類識別數據集，比MINIST更復雜，而且是RGB彩色圖片。

看看較簡單的LeNet-5可以達到多少准確率。網絡結構基本和前面MINIST代碼中的差不多，主要是輸入圖片的通道數不同，代碼如下：

# -*- coding:utf-8 -*-

u"""LeNet卷積神經網絡訓練學習CIFAR10"""

__author__ = 'zhengbiqing [e-mail]'


import torch as t
import torchvision as tv
import torch.nn as nn
import torch.optim as optim
import torchvision.transforms as transforms
from torchvision.transforms import ToPILImage
import torch.backends.cudnn as cudnn

import datetime
import argparse


# 樣本讀取線程數
WORKERS = 4

# 網絡參賽保存文件名
PARAS_FN = 'cifar_lenet_params.pkl'

# minist數據存放位置
ROOT = '/home/zbq/PycharmProjects/cifar'

# 目標函數
loss_func = nn.CrossEntropyLoss()

# 最優結果
best_acc = 0


# 定義網絡模型
class LeNet(nn.Module):
    def __init__(self):
        super(LeNet, self).__init__()

        # 卷積層
        self.cnn = nn.Sequential(
            # 卷積層1，3通道輸入，6個卷積核，核大小5*5
            # 經過該層圖像大小變為32-5+1，28*28
            # 經2*2最大池化，圖像變為14*14
            nn.Conv2d(3, 6, 5),
            nn.ReLU(),
            nn.MaxPool2d(2),

            # 卷積層2，6輸入通道，16個卷積核，核大小5*5
            # 經過該層圖像變為14-5+1，10*10
            # 經2*2最大池化，圖像變為5*5
            nn.Conv2d(6, 16, 5),
            nn.ReLU(),
            nn.MaxPool2d(2)
        )

        # 全連接層
        self.fc = nn.Sequential(
            # 16個feature，每個feature 5*5
            nn.Linear(16 * 5 * 5, 120),
            nn.ReLU(),
            nn.Linear(120, 84),
            nn.ReLU(),
            nn.Linear(84, 10)
        )

    def forward(self, x):
        x = self.cnn(x)

        # x.size()[0]: batch size
        x = x.view(x.size()[0], -1)
        x = self.fc(x)

        return x


'''
訓練並測試網絡
net：網絡模型
train_data_load：訓練數據集
optimizer：優化器
epoch：第幾次訓練迭代
log_interval：訓練過程中損失函數值和准確率的打印頻率
'''
def net_train(net, train_data_load, optimizer, epoch, log_interval):
    net.train()

    begin = datetime.datetime.now()

    # 樣本總數
    total = len(train_data_load.dataset)

    # 樣本批次訓練的損失函數值的和
    train_loss = 0

    # 識別正確的樣本數
    ok = 0

    for i, data in enumerate(train_data_load, 0):
        img, label = data
        img, label = img.cuda(), label.cuda()

        optimizer.zero_grad()

        outs = net(img)
        loss = loss_func(outs, label)
        loss.backward()
        optimizer.step()

        # 累加損失值和訓練樣本數
        train_loss += loss.item()
        # total += label.size(0)

        _, predicted = t.max(outs.data, 1)
        # 累加識別正確的樣本數
        ok += (predicted == label).sum()

        if (i + 1) % log_interval == 0:
            # 訓練結果輸出

            # 損失函數均值
            loss_mean = train_loss / (i + 1)

            # 已訓練的樣本數
            traind_total = (i + 1) * len(label)

            # 准確度
            acc = 100. * ok / traind_total

            # 一個迭代的進度百分比
            progress = 100. * traind_total / total

            print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}  Acc: {:.6f}'.format(
                epoch, traind_total, total, progress, loss_mean, acc))

    end = datetime.datetime.now()
    print('one epoch spend: ', end - begin)


'''
用測試集檢查准確率
'''
def net_test(net, test_data_load, epoch):
    net.eval()

    ok = 0

    for i, data in enumerate(test_data_load):
        img, label = data
        img, label = img.cuda(), label.cuda()

        outs = net(img)
        _, pre = t.max(outs.data, 1)
        ok += (pre == label).sum()

    acc = ok.item() * 100. / (len(test_data_load.dataset))
    print('EPOCH:{}, ACC:{}\n'.format(epoch, acc))

    global best_acc
    if acc > best_acc:
        best_acc = acc


'''
顯示數據集中一個圖片
'''
def img_show(dataset, index):
    classes = ('plane', 'car', 'bird', 'cat',
               'deer', 'dog', 'frog', 'horse', 'ship', 'truck')

    show = ToPILImage()

    data, label = dataset[index]
    print('img is a ', classes[label])
    show((data + 1) / 2).resize((100, 100)).show()


def main():
    # 訓練超參數設置，可通過命令行設置
    parser = argparse.ArgumentParser(description='PyTorch CIFA10 LeNet Example')
    parser.add_argument('--batch-size', type=int, default=64, metavar='N',
                        help='input batch size for training (default: 64)')
    parser.add_argument('--test-batch-size', type=int, default=1000, metavar='N',
                        help='input batch size for testing (default: 1000)')
    parser.add_argument('--epochs', type=int, default=20, metavar='N',
                        help='number of epochs to train (default: 20)')
    parser.add_argument('--lr', type=float, default=0.01, metavar='LR',
                        help='learning rate (default: 0.01)')
    parser.add_argument('--momentum', type=float, default=0.9, metavar='M',
                        help='SGD momentum (default: 0.9)')
    parser.add_argument('--log-interval', type=int, default=100, metavar='N',
                        help='how many batches to wait before logging training status (default: 100)')
    parser.add_argument('--no-train', action='store_true', default=False,
                        help='If train the Model')
    parser.add_argument('--save-model', action='store_true', default=False,
                        help='For Saving the current Model')
    args = parser.parse_args()

    # 圖像數值轉換，ToTensor源碼注釋
    """Convert a ``PIL Image`` or ``numpy.ndarray`` to tensor.
        Converts a PIL Image or numpy.ndarray (H x W x C) in the range
        [0, 255] to a torch.FloatTensor of shape (C x H x W) in the range [0.0, 1.0].
        """
    # 歸一化把[0.0, 1.0]變換為[-1,1], ([0, 1] - 0.5) / 0.5 = [-1, 1]
    transform = tv.transforms.Compose([
        transforms.ToTensor(),
        transforms.Normalize([0.5, 0.5, 0.5], [0.5, 0.5, 0.5])])

    # 定義數據集
    train_data = tv.datasets.CIFAR10(root=ROOT, train=True, download=True, transform=transform)
    test_data = tv.datasets.CIFAR10(root=ROOT, train=False, download=False, transform=transform)

    train_load = t.utils.data.DataLoader(train_data, batch_size=args.batch_size, shuffle=True, num_workers=WORKERS)
    test_load = t.utils.data.DataLoader(test_data, batch_size=args.test_batch_size, shuffle=False, num_workers=WORKERS)

    net = LeNet().cuda()
    print(net)

    # 如果不訓練，直接加載保存的網絡參數進行測試集驗證
    if args.no_train:
        net.load_state_dict(t.load(PARAS_FN))
        net_test(net, test_load, 0)
        return

    optimizer = optim.SGD(net.parameters(), lr=args.lr, momentum=args.momentum)

    start_time = datetime.datetime.now()

    for epoch in range(1, args.epochs + 1):
        net_train(net, train_load, optimizer, epoch, args.log_interval)

        # 每個epoch結束后用測試集檢查識別准確度
        net_test(net, test_load, epoch)

    end_time = datetime.datetime.now()

    global best_acc
    print('CIFAR10 pytorch LeNet Train: EPOCH:{}, BATCH_SZ:{}, LR:{}, ACC:{}'.format(args.epochs, args.batch_size, args.lr, best_acc))
    print('train spend time: ', end_time - start_time)

    if args.save_model:
        t.save(net.state_dict(), PARAS_FN)


if __name__ == '__main__':
    main()

 

運行結果如下：

Files already downloaded and verified
LeNet(
  (cnn): Sequential(
    (0): Conv2d(3, 6, kernel_size=(5, 5), stride=(1, 1))
    (1): ReLU()
    (2): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (3): Conv2d(6, 16, kernel_size=(5, 5), stride=(1, 1))
    (4): ReLU()
    (5): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  )
  (fc): Sequential(
    (0): Linear(in_features=400, out_features=120, bias=True)
    (1): ReLU()
    (2): Linear(in_features=120, out_features=84, bias=True)
    (3): ReLU()
    (4): Linear(in_features=84, out_features=10, bias=True)
  )
)
Train Epoch: 1 [6400/50000 (13%)]    Loss: 2.297558  Acc: 10.000000
Train Epoch: 1 [12800/50000 (26%)]    Loss: 2.219855  Acc: 16.000000
Train Epoch: 1 [19200/50000 (38%)]    Loss: 2.117518  Acc: 20.000000
Train Epoch: 1 [25600/50000 (51%)]    Loss: 2.030452  Acc: 23.000000
Train Epoch: 1 [32000/50000 (64%)]    Loss: 1.956154  Acc: 26.000000
Train Epoch: 1 [38400/50000 (77%)]    Loss: 1.894052  Acc: 29.000000
Train Epoch: 1 [44800/50000 (90%)]    Loss: 1.845520  Acc: 31.000000
one epoch spend:  0:00:02.007186
EPOCH:1, ACC:43.86

Train Epoch: 2 [6400/50000 (13%)]    Loss: 1.497962  Acc: 44.000000
Train Epoch: 2 [12800/50000 (26%)]    Loss: 1.471271  Acc: 45.000000
Train Epoch: 2 [19200/50000 (38%)]    Loss: 1.458853  Acc: 46.000000
Train Epoch: 2 [25600/50000 (51%)]    Loss: 1.445787  Acc: 47.000000
Train Epoch: 2 [32000/50000 (64%)]    Loss: 1.436431  Acc: 47.000000
Train Epoch: 2 [38400/50000 (77%)]    Loss: 1.425798  Acc: 47.000000
Train Epoch: 2 [44800/50000 (90%)]    Loss: 1.415501  Acc: 48.000000
one epoch spend:  0:00:01.879316
EPOCH:2, ACC:53.16

Train Epoch: 3 [6400/50000 (13%)]    Loss: 1.288907  Acc: 52.000000
Train Epoch: 3 [12800/50000 (26%)]    Loss: 1.293646  Acc: 53.000000
Train Epoch: 3 [19200/50000 (38%)]    Loss: 1.284784  Acc: 53.000000
Train Epoch: 3 [25600/50000 (51%)]    Loss: 1.281050  Acc: 53.000000
Train Epoch: 3 [32000/50000 (64%)]    Loss: 1.281222  Acc: 53.000000
Train Epoch: 3 [38400/50000 (77%)]    Loss: 1.269620  Acc: 54.000000
Train Epoch: 3 [44800/50000 (90%)]    Loss: 1.262982  Acc: 54.000000
one epoch spend:  0:00:01.928787
EPOCH:3, ACC:54.31

Train Epoch: 4 [6400/50000 (13%)]    Loss: 1.157912  Acc: 58.000000
Train Epoch: 4 [12800/50000 (26%)]    Loss: 1.157038  Acc: 58.000000
Train Epoch: 4 [19200/50000 (38%)]    Loss: 1.164880  Acc: 58.000000
Train Epoch: 4 [25600/50000 (51%)]    Loss: 1.169460  Acc: 58.000000
Train Epoch: 4 [32000/50000 (64%)]    Loss: 1.169655  Acc: 58.000000
Train Epoch: 4 [38400/50000 (77%)]    Loss: 1.169239  Acc: 58.000000
Train Epoch: 4 [44800/50000 (90%)]    Loss: 1.159252  Acc: 58.000000
one epoch spend:  0:00:01.928551
EPOCH:4, ACC:60.15

Train Epoch: 5 [6400/50000 (13%)]    Loss: 1.081296  Acc: 61.000000
Train Epoch: 5 [12800/50000 (26%)]    Loss: 1.073868  Acc: 61.000000
Train Epoch: 5 [19200/50000 (38%)]    Loss: 1.086076  Acc: 61.000000
Train Epoch: 5 [25600/50000 (51%)]    Loss: 1.088019  Acc: 61.000000
Train Epoch: 5 [32000/50000 (64%)]    Loss: 1.083983  Acc: 61.000000
Train Epoch: 5 [38400/50000 (77%)]    Loss: 1.088050  Acc: 61.000000
Train Epoch: 5 [44800/50000 (90%)]    Loss: 1.087298  Acc: 61.000000
one epoch spend:  0:00:01.898825
EPOCH:5, ACC:59.84

Train Epoch: 6 [6400/50000 (13%)]    Loss: 0.979352  Acc: 65.000000
Train Epoch: 6 [12800/50000 (26%)]    Loss: 1.005338  Acc: 64.000000
Train Epoch: 6 [19200/50000 (38%)]    Loss: 1.019300  Acc: 63.000000
Train Epoch: 6 [25600/50000 (51%)]    Loss: 1.022704  Acc: 63.000000
Train Epoch: 6 [32000/50000 (64%)]    Loss: 1.021217  Acc: 63.000000
Train Epoch: 6 [38400/50000 (77%)]    Loss: 1.022035  Acc: 63.000000
Train Epoch: 6 [44800/50000 (90%)]    Loss: 1.024987  Acc: 63.000000
one epoch spend:  0:00:01.926922
EPOCH:6, ACC:60.04

Train Epoch: 7 [6400/50000 (13%)]    Loss: 0.952975  Acc: 66.000000
Train Epoch: 7 [12800/50000 (26%)]    Loss: 0.965437  Acc: 65.000000
Train Epoch: 7 [19200/50000 (38%)]    Loss: 0.964711  Acc: 65.000000
Train Epoch: 7 [25600/50000 (51%)]    Loss: 0.962520  Acc: 65.000000
Train Epoch: 7 [32000/50000 (64%)]    Loss: 0.964768  Acc: 65.000000
Train Epoch: 7 [38400/50000 (77%)]    Loss: 0.966530  Acc: 65.000000
Train Epoch: 7 [44800/50000 (90%)]    Loss: 0.971995  Acc: 65.000000
one epoch spend:  0:00:01.858537
EPOCH:7, ACC:62.63

Train Epoch: 8 [6400/50000 (13%)]    Loss: 0.901441  Acc: 67.000000
Train Epoch: 8 [12800/50000 (26%)]    Loss: 0.896776  Acc: 68.000000
Train Epoch: 8 [19200/50000 (38%)]    Loss: 0.898365  Acc: 68.000000
Train Epoch: 8 [25600/50000 (51%)]    Loss: 0.898383  Acc: 68.000000
Train Epoch: 8 [32000/50000 (64%)]    Loss: 0.909455  Acc: 67.000000
Train Epoch: 8 [38400/50000 (77%)]    Loss: 0.910068  Acc: 67.000000
Train Epoch: 8 [44800/50000 (90%)]    Loss: 0.914733  Acc: 67.000000
one epoch spend:  0:00:01.849259
EPOCH:8, ACC:62.99

Train Epoch: 9 [6400/50000 (13%)]    Loss: 0.842184  Acc: 69.000000
Train Epoch: 9 [12800/50000 (26%)]    Loss: 0.853178  Acc: 69.000000
Train Epoch: 9 [19200/50000 (38%)]    Loss: 0.863828  Acc: 69.000000
Train Epoch: 9 [25600/50000 (51%)]    Loss: 0.868452  Acc: 69.000000
Train Epoch: 9 [32000/50000 (64%)]    Loss: 0.870991  Acc: 69.000000
Train Epoch: 9 [38400/50000 (77%)]    Loss: 0.874963  Acc: 69.000000
Train Epoch: 9 [44800/50000 (90%)]    Loss: 0.878533  Acc: 68.000000
one epoch spend:  0:00:01.954615
EPOCH:9, ACC:62.5

Train Epoch: 10 [6400/50000 (13%)]    Loss: 0.837819  Acc: 70.000000
Train Epoch: 10 [12800/50000 (26%)]    Loss: 0.823905  Acc: 70.000000
Train Epoch: 10 [19200/50000 (38%)]    Loss: 0.833733  Acc: 70.000000
Train Epoch: 10 [25600/50000 (51%)]    Loss: 0.838861  Acc: 70.000000
Train Epoch: 10 [32000/50000 (64%)]    Loss: 0.841117  Acc: 70.000000
Train Epoch: 10 [38400/50000 (77%)]    Loss: 0.849762  Acc: 69.000000
Train Epoch: 10 [44800/50000 (90%)]    Loss: 0.850071  Acc: 69.000000
one epoch spend:  0:00:01.812348
EPOCH:10, ACC:63.48

Train Epoch: 11 [6400/50000 (13%)]    Loss: 0.781857  Acc: 72.000000
Train Epoch: 11 [12800/50000 (26%)]    Loss: 0.773329  Acc: 72.000000
Train Epoch: 11 [19200/50000 (38%)]    Loss: 0.785191  Acc: 72.000000
Train Epoch: 11 [25600/50000 (51%)]    Loss: 0.797921  Acc: 71.000000
Train Epoch: 11 [32000/50000 (64%)]    Loss: 0.802146  Acc: 71.000000
Train Epoch: 11 [38400/50000 (77%)]    Loss: 0.804404  Acc: 71.000000
Train Epoch: 11 [44800/50000 (90%)]    Loss: 0.805919  Acc: 71.000000
one epoch spend:  0:00:01.881838
EPOCH:11, ACC:63.72

Train Epoch: 12 [6400/50000 (13%)]    Loss: 0.734165  Acc: 74.000000
Train Epoch: 12 [12800/50000 (26%)]    Loss: 0.739923  Acc: 74.000000
Train Epoch: 12 [19200/50000 (38%)]    Loss: 0.753080  Acc: 73.000000
Train Epoch: 12 [25600/50000 (51%)]    Loss: 0.755026  Acc: 73.000000
Train Epoch: 12 [32000/50000 (64%)]    Loss: 0.758760  Acc: 73.000000
Train Epoch: 12 [38400/50000 (77%)]    Loss: 0.765208  Acc: 72.000000
Train Epoch: 12 [44800/50000 (90%)]    Loss: 0.774539  Acc: 72.000000
one epoch spend:  0:00:01.856290
EPOCH:12, ACC:63.71

Train Epoch: 13 [6400/50000 (13%)]    Loss: 0.709528  Acc: 75.000000
Train Epoch: 13 [12800/50000 (26%)]    Loss: 0.713831  Acc: 74.000000
Train Epoch: 13 [19200/50000 (38%)]    Loss: 0.720146  Acc: 74.000000
Train Epoch: 13 [25600/50000 (51%)]    Loss: 0.723680  Acc: 74.000000
Train Epoch: 13 [32000/50000 (64%)]    Loss: 0.730473  Acc: 73.000000
Train Epoch: 13 [38400/50000 (77%)]    Loss: 0.742575  Acc: 73.000000
Train Epoch: 13 [44800/50000 (90%)]    Loss: 0.744857  Acc: 73.000000
one epoch spend:  0:00:01.808256
EPOCH:13, ACC:61.71

Train Epoch: 14 [6400/50000 (13%)]    Loss: 0.700821  Acc: 74.000000
Train Epoch: 14 [12800/50000 (26%)]    Loss: 0.691082  Acc: 75.000000
Train Epoch: 14 [19200/50000 (38%)]    Loss: 0.693119  Acc: 75.000000
Train Epoch: 14 [25600/50000 (51%)]    Loss: 0.706147  Acc: 74.000000
Train Epoch: 14 [32000/50000 (64%)]    Loss: 0.710033  Acc: 74.000000
Train Epoch: 14 [38400/50000 (77%)]    Loss: 0.717097  Acc: 74.000000
Train Epoch: 14 [44800/50000 (90%)]    Loss: 0.724987  Acc: 74.000000
one epoch spend:  0:00:01.797417
EPOCH:14, ACC:63.15

Train Epoch: 15 [6400/50000 (13%)]    Loss: 0.624073  Acc: 77.000000
Train Epoch: 15 [12800/50000 (26%)]    Loss: 0.637354  Acc: 77.000000
Train Epoch: 15 [19200/50000 (38%)]    Loss: 0.646385  Acc: 76.000000
Train Epoch: 15 [25600/50000 (51%)]    Loss: 0.662080  Acc: 76.000000
Train Epoch: 15 [32000/50000 (64%)]    Loss: 0.668658  Acc: 76.000000
Train Epoch: 15 [38400/50000 (77%)]    Loss: 0.679682  Acc: 75.000000
Train Epoch: 15 [44800/50000 (90%)]    Loss: 0.688876  Acc: 75.000000
one epoch spend:  0:00:01.916400
EPOCH:15, ACC:62.81

Train Epoch: 16 [6400/50000 (13%)]    Loss: 0.611007  Acc: 78.000000
Train Epoch: 16 [12800/50000 (26%)]    Loss: 0.612629  Acc: 78.000000
Train Epoch: 16 [19200/50000 (38%)]    Loss: 0.622980  Acc: 77.000000
Train Epoch: 16 [25600/50000 (51%)]    Loss: 0.638267  Acc: 77.000000
Train Epoch: 16 [32000/50000 (64%)]    Loss: 0.650756  Acc: 76.000000
Train Epoch: 16 [38400/50000 (77%)]    Loss: 0.656675  Acc: 76.000000
Train Epoch: 16 [44800/50000 (90%)]    Loss: 0.665181  Acc: 75.000000
one epoch spend:  0:00:01.878367
EPOCH:16, ACC:61.64

Train Epoch: 17 [6400/50000 (13%)]    Loss: 0.591583  Acc: 78.000000
Train Epoch: 17 [12800/50000 (26%)]    Loss: 0.601943  Acc: 78.000000
Train Epoch: 17 [19200/50000 (38%)]    Loss: 0.612084  Acc: 78.000000
Train Epoch: 17 [25600/50000 (51%)]    Loss: 0.619225  Acc: 77.000000
Train Epoch: 17 [32000/50000 (64%)]    Loss: 0.633562  Acc: 77.000000
Train Epoch: 17 [38400/50000 (77%)]    Loss: 0.641217  Acc: 77.000000
Train Epoch: 17 [44800/50000 (90%)]    Loss: 0.648393  Acc: 76.000000
one epoch spend:  0:00:01.894760
EPOCH:17, ACC:61.44

Train Epoch: 18 [6400/50000 (13%)]    Loss: 0.553651  Acc: 80.000000
Train Epoch: 18 [12800/50000 (26%)]    Loss: 0.569668  Acc: 79.000000
Train Epoch: 18 [19200/50000 (38%)]    Loss: 0.584057  Acc: 78.000000
Train Epoch: 18 [25600/50000 (51%)]    Loss: 0.598776  Acc: 78.000000
Train Epoch: 18 [32000/50000 (64%)]    Loss: 0.610767  Acc: 78.000000
Train Epoch: 18 [38400/50000 (77%)]    Loss: 0.617563  Acc: 77.000000
Train Epoch: 18 [44800/50000 (90%)]    Loss: 0.628669  Acc: 77.000000
one epoch spend:  0:00:01.925175
EPOCH:18, ACC:62.46

Train Epoch: 19 [6400/50000 (13%)]    Loss: 0.554530  Acc: 79.000000
Train Epoch: 19 [12800/50000 (26%)]    Loss: 0.574952  Acc: 78.000000
Train Epoch: 19 [19200/50000 (38%)]    Loss: 0.576819  Acc: 79.000000
Train Epoch: 19 [25600/50000 (51%)]    Loss: 0.584052  Acc: 78.000000
Train Epoch: 19 [32000/50000 (64%)]    Loss: 0.590673  Acc: 78.000000
Train Epoch: 19 [38400/50000 (77%)]    Loss: 0.599807  Acc: 78.000000
Train Epoch: 19 [44800/50000 (90%)]    Loss: 0.607849  Acc: 78.000000
one epoch spend:  0:00:01.827582
EPOCH:19, ACC:62.16

Train Epoch: 20 [6400/50000 (13%)]    Loss: 0.534505  Acc: 80.000000
Train Epoch: 20 [12800/50000 (26%)]    Loss: 0.547133  Acc: 80.000000
Train Epoch: 20 [19200/50000 (38%)]    Loss: 0.557482  Acc: 79.000000
Train Epoch: 20 [25600/50000 (51%)]    Loss: 0.567949  Acc: 79.000000
Train Epoch: 20 [32000/50000 (64%)]    Loss: 0.579047  Acc: 79.000000
Train Epoch: 20 [38400/50000 (77%)]    Loss: 0.591825  Acc: 78.000000
Train Epoch: 20 [44800/50000 (90%)]    Loss: 0.598099  Acc: 78.000000
one epoch spend:  0:00:01.846124
EPOCH:20, ACC:62.47

CIFAR10 pytorch LeNet Train: EPOCH:20, BATCH_SZ:64, LR:0.01, ACC:63.72
train spend time:  0:00:46.669295

Process finished with exit code 0

訓練的lenet准確度在63%左右，遠低於MINIST的99%，簡單的LeNet對較復雜的圖片准確度不高。