def sdn_loss(output, label, coeffs=None):
total_loss = 0.0
if coeffs is None:
coeffs = [1 for _ in range(len(output) - 1)]
for ic_id in range(len(output) - 1):
total_loss += float(coeffs[ic_id]) * af.get_loss_criterion()(
output[ic_id], label)
total_loss += af.get_loss_criterion()(output[-1], label)
return total_loss
def sdn_training_step(optimizer, model, coeffs, batch, device):
b_x = batch[0].to(device)
b_y = batch[1].to(device)
output = model(b_x)
optimizer.zero_grad() #clear gradients for this training step
total_loss = 0.0
for ic_id in range(model.num_output - 1):
cur_output = output[ic_id]
cur_loss = float(coeffs[ic_id])*af.get_loss_criterion()(cur_output, b_y)
total_loss += cur_loss
total_loss += af.get_loss_criterion()(output[-1], b_y)
total_loss.backward()
optimizer.step() # apply gradients
return total_loss
def cnn_training_step(model, optimizer, data, labels, device='cpu'):
b_x = data.to(device) # batch x
b_y = labels.to(device) # batch y
output = model(b_x) # cnn final output
criterion = af.get_loss_criterion()
loss = criterion(output, b_y) # cross entropy loss
optimizer.zero_grad() # clear gradients for this training step
loss.backward() # backpropagation, compute gradients
optimizer.step() # apply gradients
def sdn_ic_only_step(optimizer, model, batch, device):
b_x = batch[0].to(device)
b_y = batch[1].to(device)
output = model(b_x)
optimizer.zero_grad() #clear gradients for this training step
total_loss = 0.0
for output_id, cur_output in enumerate(output):
if output_id == model.num_output - 1: # last output
break
cur_loss = af.get_loss_criterion()(cur_output, b_y)
total_loss += cur_loss
total_loss.backward()
optimizer.step() # apply gradients
return total_loss