def Testing(pred, y, name):
accuracy_test = score_py3.accuracy(pred, y.data)
f1_test = score_py3.score_f1(pred, y.data)
if Args.show_log:
print('Epoch: %s | %s Train Accuracy: %.5f | Train F1: %.5f' %
(epoch, name, accuracy_test, f1_test))
return f1_test
def training(output, loss, optimizer, y, name, Args):
optimizer.zero_grad() # clear gradients for backward
loss.backward() # backpropagation, compute gradients
optimizer.step() # apply gradients
if Args.show_log:
if step % 1 == 0:
if Args.cuda:
pred = torch.max(output, 1)[1].cuda().data.squeeze()
else:
pred = torch.max(output, 1)[1].data.squeeze()
# accuracy
accuracy_train = score_py3.accuracy(pred, y.data)
# F1
f1_train = score_py3.score_f1(pred, y.data)
# print
print(
'Epoch: %s | %s Train Accuracy: %.5f | Train F1: %.5f | Loss: %.2f'
% (epoch, name, accuracy_train, f1_train, loss.data))
output = vgg16(x)
loss = loss_func(output, y) # get loss
optimizer.zero_grad() # clear gradients for backward
loss.backward() # backpropagation, compute gradients
optimizer.step() # apply gradients
##### train evaluate #####
if Args.show_log:
if step % 1 == 0:
if Args.cuda:
pred = torch.max(output, 1)[1].cuda().data.squeeze()
else:
pred = torch.max(output, 1)[1].data.squeeze()
# accuracy
accuracy_train = score_py3.accuracy(pred, y.data)
# F1
f1_train = score_py3.score_f1(pred, y.data)
# print
print(
'Epoch: %s | Train Accuracy: %.5f | Train F1: %.5f | Loss: %.2f'
% (epoch, accuracy_train, f1_train, loss.data))
##### Test #####
all_y = []
all_pred = []
for step, (x, y) in enumerate(loader_test):
if Args.cuda:
x = x.cuda()
y = y.cuda()
vgg16.eval()
output = vgg16(x)
if Args.cuda:
pred = torch.max(output, 1)[1].cuda().data.squeeze()