基于Pytorch深度學習神經網絡MNIST手寫數字識別系統源碼（帶界面和手寫畫板）

?第一步：準備數據

mnist開源數據集

第二步：搭建模型

我們這里搭建了一個LeNet5網絡

參考代碼如下：

import torch
from torch import nnclass Reshape(nn.Module):def forward(self, x):return x.view(-1, 1, 28, 28)class LeNet5(nn.Module):def __init__(self):super(LeNet5, self).__init__()self.net = nn.Sequential(Reshape(),# CONV1, ReLU1, POOL1nn.Conv2d(in_channels=1, out_channels=6, kernel_size=5, padding=2),# nn.Conv2d(in_channels=1, out_channels=6, kernel_size=5),nn.ReLU(),nn.MaxPool2d(kernel_size=2, stride=2),# CONV2, ReLU2, POOL2nn.Conv2d(in_channels=6, out_channels=16, kernel_size=5),nn.ReLU(),nn.MaxPool2d(kernel_size=2, stride=2),nn.Flatten(),# FC1nn.Linear(in_features=16 * 5 * 5, out_features=120),nn.ReLU(),# FC2nn.Linear(in_features=120, out_features=84),nn.ReLU(),# FC3nn.Linear(in_features=84, out_features=10))# 添加softmax層self.softmax = nn.Softmax()def forward(self, x):logits = self.net(x)# 將logits轉為概率prob = self.softmax(logits)return probif __name__ == '__main__':model = LeNet5()X = torch.rand(size=(256, 1, 28, 28), dtype=torch.float32)for layer in model.net:X = layer(X)print(layer.__class__.__name__, '\toutput shape: \t', X.shape)X = torch.rand(size=(1, 1, 28, 28), dtype=torch.float32)print(model(X))

第三步：訓練代碼

import torch
from torch import nn
from torchvision import datasets
from torchvision.transforms import ToTensor
from torch.utils.data import DataLoaderfrom model import LeNet5# DATASET
train_data = datasets.MNIST(root='./data',train=False,download=True,transform=ToTensor()
)test_data = datasets.MNIST(root='./data',train=False,download=True,transform=ToTensor()
)# PREPROCESS
batch_size = 256
train_dataloader = DataLoader(dataset=train_data, batch_size=batch_size)
test_dataloader = DataLoader(dataset=test_data, batch_size=batch_size)
for X, y in train_dataloader:print(X.shape)		# torch.Size([256, 1, 28, 28])print(y.shape)		# torch.Size([256])break# MODEL
device = 'cuda' if torch.cuda.is_available() else 'cpu'
model = LeNet5().to(device)# TRAIN MODEL
loss_func = nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(params=model.parameters())def train(dataloader, model, loss_func, optimizer, epoch):model.train()data_size = len(dataloader.dataset)for batch, (X, y) in enumerate(dataloader):X, y = X.to(device), y.to(device)y_hat = model(X)loss = loss_func(y_hat, y)optimizer.zero_grad()loss.backward()optimizer.step()loss, current = loss.item(), batch * len(X)print(f'EPOCH{epoch+1}\tloss: {loss:>7f}', end='\t')# Test model
def test(dataloader, model, loss_fn):size = len(dataloader.dataset)num_batches = len(dataloader)model.eval()test_loss, correct = 0, 0with torch.no_grad():for X, y in dataloader:X, y = X.to(device), y.to(device)pred = model(X)test_loss += loss_fn(pred, y).item()correct += (pred.argmax(1) == y).type(torch.float).sum().item()test_loss /= num_batchescorrect /= sizeprint(f'Test Error: Accuracy: {(100 * correct):>0.1f}%, Average loss: {test_loss:>8f}\n')if __name__ == '__main__':epoches = 80for epoch in range(epoches):train(train_dataloader, model, loss_func, optimizer, epoch)test(test_dataloader, model, loss_func)# Save modelstorch.save(model.state_dict(), 'model.pth')print('Saved PyTorch LeNet5 State to model.pth')

第四步：統計訓練過程

EPOCH1	loss: 1.908403	Test Error: Accuracy: 58.3%, Average loss: 1.943602EPOCH2	loss: 1.776060	Test Error: Accuracy: 72.2%, Average loss: 1.750917EPOCH3	loss: 1.717706	Test Error: Accuracy: 73.6%, Average loss: 1.730332EPOCH4	loss: 1.719344	Test Error: Accuracy: 76.0%, Average loss: 1.703456EPOCH5	loss: 1.659312	Test Error: Accuracy: 76.6%, Average loss: 1.694500EPOCH6	loss: 1.647946	Test Error: Accuracy: 76.9%, Average loss: 1.691286EPOCH7	loss: 1.653712	Test Error: Accuracy: 77.0%, Average loss: 1.690819EPOCH8	loss: 1.653270	Test Error: Accuracy: 76.8%, Average loss: 1.692459EPOCH9	loss: 1.649021	Test Error: Accuracy: 77.5%, Average loss: 1.686158EPOCH10	loss: 1.648204	Test Error: Accuracy: 78.3%, Average loss: 1.678802EPOCH11	loss: 1.647159	Test Error: Accuracy: 78.4%, Average loss: 1.676133EPOCH12	loss: 1.647390	Test Error: Accuracy: 78.6%, Average loss: 1.674455EPOCH13	loss: 1.646807	Test Error: Accuracy: 78.4%, Average loss: 1.675752EPOCH14	loss: 1.630824	Test Error: Accuracy: 79.1%, Average loss: 1.668470EPOCH15	loss: 1.524222	Test Error: Accuracy: 86.3%, Average loss: 1.599240EPOCH16	loss: 1.524022	Test Error: Accuracy: 86.7%, Average loss: 1.594947EPOCH17	loss: 1.524296	Test Error: Accuracy: 87.1%, Average loss: 1.588946EPOCH18	loss: 1.523599	Test Error: Accuracy: 87.3%, Average loss: 1.588275EPOCH19	loss: 1.523655	Test Error: Accuracy: 87.5%, Average loss: 1.586576EPOCH20	loss: 1.523659	Test Error: Accuracy: 88.2%, Average loss: 1.579286EPOCH21	loss: 1.523733	Test Error: Accuracy: 87.9%, Average loss: 1.582472EPOCH22	loss: 1.523748	Test Error: Accuracy: 88.2%, Average loss: 1.578699EPOCH23	loss: 1.523788	Test Error: Accuracy: 88.0%, Average loss: 1.579700EPOCH24	loss: 1.523708	Test Error: Accuracy: 88.1%, Average loss: 1.579758EPOCH25	loss: 1.523683	Test Error: Accuracy: 88.4%, Average loss: 1.575913EPOCH26	loss: 1.523646	Test Error: Accuracy: 88.7%, Average loss: 1.572831EPOCH27	loss: 1.523654	Test Error: Accuracy: 88.9%, Average loss: 1.570528EPOCH28	loss: 1.523642	Test Error: Accuracy: 89.0%, Average loss: 1.570223EPOCH29	loss: 1.523663	Test Error: Accuracy: 89.0%, Average loss: 1.570385EPOCH30	loss: 1.523658	Test Error: Accuracy: 88.9%, Average loss: 1.571195EPOCH31	loss: 1.523653	Test Error: Accuracy: 88.4%, Average loss: 1.575981EPOCH32	loss: 1.523653	Test Error: Accuracy: 89.0%, Average loss: 1.570087EPOCH33	loss: 1.523642	Test Error: Accuracy: 88.9%, Average loss: 1.571018EPOCH34	loss: 1.523649	Test Error: Accuracy: 89.0%, Average loss: 1.570439EPOCH35	loss: 1.523629	Test Error: Accuracy: 90.4%, Average loss: 1.555473EPOCH36	loss: 1.461187	Test Error: Accuracy: 97.1%, Average loss: 1.491042EPOCH37	loss: 1.461230	Test Error: Accuracy: 97.7%, Average loss: 1.485049EPOCH38	loss: 1.461184	Test Error: Accuracy: 97.7%, Average loss: 1.485653EPOCH39	loss: 1.461156	Test Error: Accuracy: 98.2%, Average loss: 1.479966EPOCH40	loss: 1.461335	Test Error: Accuracy: 98.2%, Average loss: 1.479197EPOCH41	loss: 1.461152	Test Error: Accuracy: 98.7%, Average loss: 1.475477EPOCH42	loss: 1.461153	Test Error: Accuracy: 98.7%, Average loss: 1.475124EPOCH43	loss: 1.461153	Test Error: Accuracy: 98.9%, Average loss: 1.472885EPOCH44	loss: 1.461151	Test Error: Accuracy: 99.1%, Average loss: 1.470957EPOCH45	loss: 1.461156	Test Error: Accuracy: 99.1%, Average loss: 1.471141EPOCH46	loss: 1.461152	Test Error: Accuracy: 99.1%, Average loss: 1.470793EPOCH47	loss: 1.461151	Test Error: Accuracy: 98.8%, Average loss: 1.474548EPOCH48	loss: 1.461151	Test Error: Accuracy: 99.1%, Average loss: 1.470666EPOCH49	loss: 1.461151	Test Error: Accuracy: 99.1%, Average loss: 1.471546EPOCH50	loss: 1.461151	Test Error: Accuracy: 99.0%, Average loss: 1.471407EPOCH51	loss: 1.461151	Test Error: Accuracy: 98.8%, Average loss: 1.473795EPOCH52	loss: 1.461164	Test Error: Accuracy: 98.2%, Average loss: 1.480009EPOCH53	loss: 1.461151	Test Error: Accuracy: 99.2%, Average loss: 1.469931EPOCH54	loss: 1.461152	Test Error: Accuracy: 99.2%, Average loss: 1.469916EPOCH55	loss: 1.461151	Test Error: Accuracy: 98.9%, Average loss: 1.472574EPOCH56	loss: 1.461151	Test Error: Accuracy: 98.6%, Average loss: 1.476035EPOCH57	loss: 1.461151	Test Error: Accuracy: 98.2%, Average loss: 1.478933EPOCH58	loss: 1.461150	Test Error: Accuracy: 99.4%, Average loss: 1.468186EPOCH59	loss: 1.461151	Test Error: Accuracy: 99.4%, Average loss: 1.467602EPOCH60	loss: 1.461151	Test Error: Accuracy: 99.1%, Average loss: 1.471206EPOCH61	loss: 1.461151	Test Error: Accuracy: 98.8%, Average loss: 1.473356EPOCH62	loss: 1.461151	Test Error: Accuracy: 99.2%, Average loss: 1.470242EPOCH63	loss: 1.461150	Test Error: Accuracy: 99.1%, Average loss: 1.470826EPOCH64	loss: 1.461151	Test Error: Accuracy: 98.7%, Average loss: 1.474476EPOCH65	loss: 1.461150	Test Error: Accuracy: 99.3%, Average loss: 1.469116EPOCH66	loss: 1.461150	Test Error: Accuracy: 99.4%, Average loss: 1.467823EPOCH67	loss: 1.461150	Test Error: Accuracy: 99.5%, Average loss: 1.466486EPOCH68	loss: 1.461152	Test Error: Accuracy: 99.3%, Average loss: 1.468688EPOCH69	loss: 1.461150	Test Error: Accuracy: 99.5%, Average loss: 1.466256EPOCH70	loss: 1.461150	Test Error: Accuracy: 99.5%, Average loss: 1.466588EPOCH71	loss: 1.461150	Test Error: Accuracy: 99.6%, Average loss: 1.465280EPOCH72	loss: 1.461150	Test Error: Accuracy: 99.4%, Average loss: 1.467110EPOCH73	loss: 1.461151	Test Error: Accuracy: 99.6%, Average loss: 1.465245EPOCH74	loss: 1.461150	Test Error: Accuracy: 99.5%, Average loss: 1.466551EPOCH75	loss: 1.461150	Test Error: Accuracy: 99.5%, Average loss: 1.466001EPOCH76	loss: 1.461150	Test Error: Accuracy: 99.3%, Average loss: 1.468074EPOCH77	loss: 1.461151	Test Error: Accuracy: 99.6%, Average loss: 1.465709EPOCH78	loss: 1.461150	Test Error: Accuracy: 99.5%, Average loss: 1.466567EPOCH79	loss: 1.461150	Test Error: Accuracy: 99.6%, Average loss: 1.464922EPOCH80	loss: 1.461150	Test Error: Accuracy: 99.6%, Average loss: 1.465109

第五步：搭建GUI界面