有条件的WGAN-GP在Keras中的实现

Question

我正在将WGAN-GP扩展为有条件的代码库，可在此处找到： https://github.com/eriklindernoren/Keras-GAN/blob/master/wgan_gp/wgan_gp.py

训练模型时，它似乎不受标签的限制。这就是我建立模型的方式。

Test = ['ASDFGH', 'QWERTYU', 'ZXCVB']
Ref = ['ASDFGY', 'QWERTYI', 'ZXCAA']

from collections import Counter

def comparer(x, y, n):
    return (len(x) == len(y)) and (sum(i != j for i, j in zip(x, y)) <= n)

res = [a for a, b in zip(Ref, Test) if comparer(a, b, 1)]

print(res)

['ASDFGY', 'QWERTYI']

绘制模型的结果为：

我不知道如何解释右侧的合并箭头。标签应串联在鉴别器中。我感觉这条线弄乱了一些东西：

    # The generator takes noise and the target label (states) as input
    # and generates the corresponding samples of that label
    noise = Input(shape=(self.latent_size, ), name="noise")
    label = Input(shape=(self.label_size, ), name="labels")
    real_samples = Input(shape=(self.input_size,), name="real")

    self.discriminator = self.build_discriminator()
    self.generator = self.build_generator([noise, label])

    # First we train the discriminator
    self.generator.trainable = False
    fake_samples = self.generator([noise, label])

    fake = self.discriminator([fake_samples, label])
    valid = self.discriminator([real_samples, label])

    interpolated = Lambda(self.random_weighted_average)([real_samples, fake_samples])
    valid_interp = self.discriminator([interpolated, label])

    self.d_model = Model([real_samples, noise, label],
                         [valid, fake, valid_interp],
                         name="discriminator")

    # Time to train the generator
    self.discriminator.trainable = False
    self.generator.trainable = True

    noise_gen = Input(shape=(self.latent_size,), name="noise_gen")

    fake_samples = self.generator([noise_gen, label])
    valid = self.discriminator([fake_samples, label])

    self.g_model = Model([noise_gen, label], valid, name="generator")
    self.g_model.compile(loss=self.wasserstein_loss, optimizer=optimizer)

由于我只是传递标签，而且我不知道Keras如何将输入路由到其他输出。