def traindata(device):
net=nn.netfunction()
criterion=nn.criterionfunction()
optimizer=nn.optimizerfunction()
trainloader=encode.trainloader
for epoch in range(3): # loop over the dataset multiple times
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.forward(inputs)
loss = criterion(outputs, labels)
loss.backward()
optimizer.step()
# print statistics
running_loss += loss.item()
if i % 200 == 199: # print every 200 mini-batches
print('[%d, %5d] loss: %.3f' %
(epoch + 1, i + 1, running_loss / 200))
running_loss = 0.0
print('Finished Training')
return net
def getforwardconvlayer():
dataiter = iter(encode.testloader)
images, labels = dataiter.next()
net = neuralnet.netfunction()
new_net = nn.Sequential(*list(net.features.children())[:-2])
outputs = new_net(images)
#print(outputs[0])
features = outputs[3][5].data.numpy()
#print(img.size())
features = features / 2 + 0.5 # convert back to image and unnormalize
plotfeatures(images[3], outputs[3])
def decode():
# get some random training images
dataiter = iter(encode.trainloader)
images, labels = dataiter.next()
# print labels
print('Actual: ',
' '.join('%5s' % encode.classes[labels[j]] for j in range(4)))
#print classifications
net = neuralnet.netfunction()
outputs = net(images)
_, predicted = torch.max(outputs, 1)
print('Predicted: ',
' '.join('%5s' % encode.classes[predicted[j]] for j in range(4)))
# show images
imshow(torchvision.utils.make_grid(images))
def dd_helper(image, layer, iterations, lr):
net = neuralnet.netfunction()
modulelist = list(net.features.modules())
input = Variable(encode.transform(image).unsqueeze(0), requires_grad=True)
net.zero_grad()
for i in range(iterations):
# print('Iteration: ', i)
out = input
for j in range(layer):
out = modulelist[j + 1](out)
loss = out.norm()
loss.backward()
input.data = input.data + lr * input.grad.data
input = input.data.squeeze()
input.transpose_(0, 1)
input.transpose_(1, 2)
input = np.clip(deprocess(input), 0, 1)
im = Image.fromarray(np.uint8(input * 255))
return im
#decode.getforwardconvlayer()
dataiter = iter(encode.testloader)
images, labels = dataiter.next()
image = torchvision.transforms.ToPILImage()(images[3])
print(image)
layer = 1
iterations = 5
lr = 0.4
octave_scale = 1
num_octaves = 1
decode.deep_dream_image(image, layer, iterations, lr, octave_scale,
num_octaves)
elif (cmdlist[0] == "save"):
torch.save(model.state_dict(), cmdlist[1])
elif (cmdlist[0] == "load"):
if (ismodel(cmdlist[1])):
model = neuralnet.netfunction()
model.load_state_dict(torch.load(cmdlist[1]))
model.eval()
elif (cmdlist[0] == "classify"):
print("classify")
elif (cmdlist[0] == "quit"):
print('end program')
exit()
else:
print("invalid command")