def main():
# Load net
cnn = CNN()
loss_func = nn.MultiLabelSoftMarginLoss()
optimizer = optim.Adam(cnn.parameters(), lr=learning_rate)
if torch.cuda.is_available():
cnn.cuda()
loss_func.cuda()
# Load data
train_dataloader = dataset.get_train_data_loader()
test_dataloader = dataset.get_test_data_loader()
# Train model
for epoch in range(num_epochs):
cnn.train()
for i, (images, labels) in enumerate(train_dataloader):
images = Variable(images)
labels = Variable(labels.long())
if torch.cuda.is_available():
images = images.cuda()
labels = labels.cuda()
predict_labels = cnn(images)
loss = loss_func(predict_labels, labels)
optimizer.zero_grad()
loss.backward()
optimizer.step()
if (i + 1) % 100 == 0:
print("epoch:", epoch, "step:", i, "loss:", loss.item())
# Save and test model
if (epoch + 1) % 10 == 0:
filename = "model" + str(epoch + 1) + ".pkl"
torch.save(cnn.state_dict(), filename)
cnn.eval()
correct = 0
total = 0
for (image, label) in test_dataloader:
vimage = Variable(image)
if torch.cuda.is_available():
vimage = vimage.cuda()
output = cnn(vimage)
predict_label = ""
for k in range(4):
predict_label += config.CHAR_SET[np.argmax(
output[0, k * config.CHAR_SET_LEN:(k + 1) *
config.CHAR_SET_LEN].data.cpu().numpy())]
true_label = one_hot.vec2text(label.numpy()[0])
total += label.size(0)
if predict_label == true_label:
correct += 1
if total % 200 == 0:
print(
'Test Accuracy of the model on the %d test images: %f %%'
% (total, 100 * correct / total))
print('Test Accuracy of the model on the %d test images: %f %%' %
(total, 100 * correct / total))
print("save and test model...")
torch.save(cnn.state_dict(), "./model.pkl") # current is model.pkl
print("save last model")
def train(model_name='model.pkl'):
cnn = CNN()
cnn.train()
print('init net')
criterion = nn.MultiLabelSoftMarginLoss()
optimizer = torch.optim.Adam(cnn.parameters(),
lr=setting.TRAIN_LEARNING_RATE)
# Train the Model
train_dataloader = dataset.get_train_data_loader()
for epoch in range(setting.TRAIN_NUM_EPOCHS):
for i, (images, labels) in enumerate(train_dataloader):
images = Variable(images)
labels = Variable(labels.float())
predict_labels = cnn(images)
loss = criterion(predict_labels, labels)
optimizer.zero_grad()
loss.backward()
optimizer.step()
print('epoch: % -3s loss: %s' % (epoch, loss.item()))
torch.save(cnn.state_dict(), model_name) # current is model.pkl
print('save last model')
# loop until all data are read
mf.input(network,n_tfiles,n_vfiles)
with tf.device('/gpu:0'):
# restore()
if(iter==0):
# Define the train computation graph
network.define_train_operations()
# Train the network
sess = tf.Session(config=tf.ConfigProto(allow_soft_placement = True)) # session with log about gpu exec
#sess= tf.Session()
try:
print(iter)
network.train(sess,iter)
iter += 1
flag = 0
# save()
except KeyboardInterrupt:
print()
flag = 1
finally:
flag = 1
sess.close()
except KeyboardInterrupt:
flag=1
except Exception as e:
flag=1
# template = "An exception of type {0} occurred. Arguments:\n{1!r}"
# message = template.format(type(e).__name__, e.args)
def train():
"""
Performs training and evaluation of MLP cnn.
NOTE: You should the cnn on the whole test set each eval_freq iterations.
"""
# YOUR TRAINING CODE GOES HERE
transform = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
])
train_data = datasets.CIFAR10('data',
train=True,
download=True,
transform=transform)
test_data = datasets.CIFAR10('data',
train=False,
download=True,
transform=transform)
train_on_gpu = torch.cuda.is_available()
num_train = len(train_data)
train_loader = torch.utils.data.DataLoader(train_data,
batch_size=FLAGS.batch_size,
shuffle=True,
num_workers=0)
test_loader = torch.utils.data.DataLoader(test_data,
batch_size=FLAGS.batch_size,
shuffle=False,
num_workers=0)
classes = [
'airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'frog',
'horse', 'ship', 'truck'
]
cnn = CNN(3, 10)
if train_on_gpu:
cnn.cuda()
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(cnn.parameters(), lr=FLAGS.learning_rate)
for epoch in range(1, FLAGS.max_steps):
class_correct = list(0. for i in range(10))
class_total = list(0. for i in range(10))
train_loss = 0.0
test_loss = 0.0
cnn.train()
for data, target in train_loader:
if train_on_gpu:
data, target = data.cuda(), target.cuda()
optimizer.zero_grad()
output = cnn(data)
loss = criterion(output, target)
loss.backward()
optimizer.step()
_, pred = torch.max(output, 1)
correct_tensor = pred.eq(target.data.view_as(pred))
train_loss += loss.item() * data.size(0)
correct = np.squeeze(
correct_tensor.numpy()) if not train_on_gpu else np.squeeze(
correct_tensor.cpu().numpy())
for i in range(len(target.data)):
label = target.data[i]
class_correct[label] += correct[i].item()
class_total[label] += 1
if epoch % FLAGS.eval_freq == 0:
test_correct = list(0. for i in range(10))
test_total = list(0. for i in range(10))
cnn.eval()
for data, target in test_loader:
if train_on_gpu:
data, target = data.cuda(), target.cuda()
output = cnn(data)
_, pred = torch.max(output, 1)
correct_tensor = pred.eq(target.data.view_as(pred))
correct = np.squeeze(correct_tensor.numpy()
) if not train_on_gpu else np.squeeze(
correct_tensor.cpu().numpy())
loss = criterion(output, target)
test_loss += loss.item() * data.size(0)
for i in range(len(target.data)):
label = target.data[i]
test_correct[label] += correct[i].item()
test_total[label] += 1
train_loss = train_loss / len(train_loader.dataset)
test_loss = test_loss / len(test_loader.dataset)
plot_epoch.append(epoch)
plot_train_loss.append(train_loss)
plot_test_loss.append(test_loss)
print(
'Epoch: {} \tTraining Loss: {:.6f} \tTest Loss: {:.6f}'.format(
epoch, train_loss, test_loss))
percent_train = accuracy(class_correct, class_total) * 100
percent_test = accuracy(test_correct, test_total) * 100
plot_train_accuracy.append(percent_train)
plot_test_accuracy.append(percent_test)
print('train accuracy: ', percent_train, 'test accuracy: ',
percent_test)
fig1 = plt.subplot(2, 1, 1)
fig2 = plt.subplot(2, 1, 2)
fig1.plot(plot_epoch,
plot_train_accuracy,
c='red',
label='training data accuracy')
fig1.plot(plot_epoch,
plot_test_accuracy,
c='blue',
label='test data accuracy')
fig1.legend()
fig2.plot(plot_epoch, plot_train_loss, c='green', label='train CE loss')
fig2.plot(plot_epoch, plot_test_loss, c='yellow', label='test CE loss')
fig2.legend()
plt.show()