KL_divergence subtraction z_dim

[lw3259111]

KL_divergence = 0.5 * tf.reduce_mean(tf.reduce_sum(tf.exp(self.enc_logvar) - self.enc_logvar + self.enc_mean**2,axis=1) - self.z_dim)
Thanks for your code, but i meet a problem for your code, In line 137, why need you substraction z_dim?
In origin pytorch code, i can't found the operator.

[abdulfatir]

In the KL expression you have -1 and you take a sum over the latent dimension. This would just sum to -z_dim which is what is written.