Question about GAN loss

[Nevermetyou65]

Hello Sir, Thanks for a very useful repository about GANs.

Would mind clarifying something about GAN loss? It's about the "sign" return by d_loss_fn and g_loss_fn in this snippet.

def get_loss_fn():
    def d_loss_fn(real_logits, fake_logits):
        return -tf.reduce_mean(tf.math.log(real_logits + 1e-10) + tf.math.log(1. - fake_logits + 1e-10))

    def g_loss_fn(fake_logits):
        return -tf.reduce_mean(tf.math.log(fake_logits + 1e-10))

    return d_loss_fn, 

Is this just a binary crossentropy? if it's just a binary crossentropy can I use the loss defined in DCGAN?
And you put a minus sign to make a positive return value am I correct?
I find that your implementation is the closest to those implemented in papers but slightly different in the sign of value.