Example code for torch is below.
I changed the code as below below.
import torch
import torch.nn as nn
import torchvision
import torchvision.transforms as transforms
import torchvision.models as models
import torch.optim as optim
# for image
import matplotlib.pyplot as plt
import numpy as np
transform = transforms.Compose([transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])
trainset = torchvision.datasets.CIFAR10(root='./data', train=True, download=True, transform=transform)
trainloader = torch.utils.data.DataLoader(trainset, batch_size=4, shuffle=True, num_workers=2)
testset = torchvision.datasets.CIFAR10(root='./data', train=False, download=True, transform=transform)
testloader = torch.utils.data.DataLoader(testset, batch_size=4, shuffle=False, num_workers=2)
classes = ('plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck')
net = models.resnet18() # torchvision 에 이미 정의된 모델을 가져옵니다.
criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(net.parameters(), lr=0.001, momentum=0.9)
print('\n===> Training Start')
device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
net = net.to(device)
if torch.cuda.device_count() > 1:
print('\n===> Training on GPU!')
net = nn.DataParallel(net)
epochs = 2 # dataset을 여러번 사용해 트레이닝을 시킵니다.
for epoch in range(epochs):
print('\n===> epoch %d' % epoch)
running_loss = 0.0
for i, data in enumerate(trainloader, 0):
# get the inputs
inputs, labels = data
inputs, labels = inputs.to(device), labels.to(device)
# zero the parameter gradients
optimizer.zero_grad()
# forward + backward + optimize
outputs = net(inputs)
loss = criterion(outputs, labels)
loss.backward()
optimizer.step()
# print statistics
running_loss += loss.item()
if i % 2000 == 1999: # print every 2000 mini-batches
print('[%d, %5d] loss: %.3f' %
(epoch + 1, i + 1, running_loss / 2000))
running_loss = 0.0
class_correct = list(0. for i in range(10))
class_total = list(0. for i in range(10))
with torch.no_grad():
for data in testloader:
images, labels = data
images, labels = images.to(device), labels.to(device)
outputs = net(images)
_, predicted = torch.max(outputs, 1) # prediction
c = (predicted == labels).squeeze()
for i in range(4):
label = labels[i]
class_correct[label] += c[i].item()
class_total[label] += 1
for i in range(10):
print('Accuracy of %5s : %2d %%' % (classes[i], 100 * class_correct[i] / class_total[i]))
I change the code as below
The changes are about batch. DataLoader.
There are pretty difference in final accuracy.
I don’t know what difference is.
import torch
import torch.nn as nn
import torchvision
import torchvision.transforms as transforms
import torchvision.models as models
import torch.optim as optim
# for image
import matplotlib.pyplot as plt
import numpy as np
transform = transforms.Compose([transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])
trainset = torchvision.datasets.CIFAR10(root='./data', train=True, download=True, transform=transform) # 50000x32x32x3
#trainloader = torch.utils.data.DataLoader(trainset, batch_size=4, shuffle=True, num_workers=2)
testset = torchvision.datasets.CIFAR10(root='./data', train=False, download=True, transform=transform) # 10000x32x32x3
testloader = torch.utils.data.DataLoader(testset, batch_size=4, shuffle=False, num_workers=2)
classes = ('plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck')
net = models.resnet18() # torchvision 에 이미 정의된 모델을 가져옵니다.
criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(net.parameters(), lr=0.001, momentum=0.9)
print('\n===> Training Start')
device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
net = net.to(device)
if torch.cuda.device_count() > 1:
print('\n===> Training on GPU!')
net = nn.DataParallel(net)
epochs = 2 # dataset을 여러번 사용해 트레이닝을 시킵니다.
batch_idx = torch.randperm(trainset.data.shape[0])
batch_size = 4
for epoch in range(epochs):
print('\n===> epoch %d' % epoch)
running_loss = 0.0
for i in range(int(trainset.data.shape[0] / batch_size)):
inputs=[]
labels=[]
mini_idx = batch_idx[i*batch_size:(i+1)*batch_size-1]
# get the inputs
for j in range(batch_size):
temp_inputs, temp_labels = trainset.__getitem__(j)
inputs.append(np.asarray(temp_inputs))
labels.append(np.asarray(temp_labels))
inputs = torch.from_numpy(np.array(inputs,dtype='float32')).float()
labels = torch.from_numpy(np.array(labels,dtype='int64')).long()
inputs, labels = inputs.to(device), labels.to(device)
# zero the parameter gradients
optimizer.zero_grad()
# forward + backward + optimize
outputs = net(inputs) # 네트워크에 입력은 [데이터수x채널x높이x너비]
loss = criterion(outputs, labels)
loss.backward()
optimizer.step()
# print statistics
running_loss += loss.item()
if i % 100 == 99: # print every 2000 mini-batches
print('[%d, %5d] loss: %.9f' %
(epoch + 1, i + 1, running_loss / 2000))
running_loss = 0.0
class_correct = list(0. for i in range(10))
class_total = list(0. for i in range(10))
with torch.no_grad():
for data in testloader:
images, labels = data
images, labels = images.to(device), labels.to(device)
outputs = net(images)
_, predicted = torch.max(outputs, 1) # prediction
c = (predicted == labels).squeeze()
for i in range(4):
label = labels[i]
class_correct[label] += c[i].item()
class_total[label] += 1
for i in range(10):
print('Accuracy of %5s : %2d %%' % (classes[i], 100 * class_correct[i] / class_total[i]))