Created by: ManWingloeng

        with fluid.program_guard(d_program):
            # declare_data(self)
            self.inputs = fluid.layers.data(shape=image_dims, name='real_images')
            self.y = fluid.layers.data(shape=[self.y_dim], name='y')
            self.z = fluid.layers.data(shape=[self.z_dim], name='z')
            self.unlabelled_inputs = fluid.layers.data(shape=image_dims, name='unlabelled_images')

            print(self.inputs)
            D_real, D_real_logits, _ = self.D(self.inputs, self.y, is_test=False)
            G_train = self.G(self.z, self.y, is_test=False)
            print(G_train)
            D_fake, D_fake_logits, _ = self.D(G_train, self.y, is_test=False, reuse=True)
            Y_c, _ = self.C(self.unlabelled_inputs, is_test=False)
            D_cla, D_cla_logits, _ = self.D(self.unlabelled_inputs, Y_c, is_test=False, reuse=True)

            ones_real = ones(D_real.shape)
            zeros_fake = zeros(D_fake.shape)
            zeros_cla = zeros(D_cla.shape)

            ce_real = fluid.layers.sigmoid_cross_entropy_with_logits(x=D_real_logits, label=ones_real)
            ce_fake = fluid.layers.sigmoid_cross_entropy_with_logits(x=D_fake_logits, label=zeros_fake)
            ce_cla = fluid.layers.sigmoid_cross_entropy_with_logits(x=D_cla_logits, label=zeros_cla)

            d_loss_real = fluid.layers.reduce_mean(ce_real)
            d_loss_fake = (1 - alpha) * fluid.layers.reduce_mean(ce_fake)
            d_loss_cla = alpha * fluid.layers.reduce_mean(ce_cla)
            self.d_loss = d_loss_real + d_loss_fake + d_loss_cla

            d_parameters = self.get_params(d_program, prefix='D')
            d_opt = fluid.optimizer.Adam(learning_rate=self.gan_lr, beta1=self.GAN_beta1)
            d_opt.minimize(self.d_loss, parameter_list=d_parameters)

ones and zeros is written as :

def ones(shape, dtype='float32'):
    return fluid.layers.ones(shape, dtype)

def zeros(shape, dtype='float32'):
    return fluid.layers.zeros(shape, dtype)

and the get_params functions is:

    def get_params(self, program, prefix):
        all_params = program.global_block().all_parameters()
        return [t.name for t in all_params if t.name.startswith(prefix)]

Anyone has point on this problem?