pérdida múltiple pytorch
b = nn.MSELoss()
a = nn.CrossEntropyLoss()
loss_a = a(output_x, x_labels)
loss_b = b(output_y, y_labels)
loss = loss_a + loss_b
loss.backward()
coder