Open sungwoong opened 6 years ago
Hi Sungwoong,
Yes, the model should work for both cases but empirically we found that using f_recon is a little bit more stable, especially on image datasets. I will take notes about it in the revised version of paper, and also in the code. Thanks very much.
Best, Trung
Thanks for your kind answer! Sungwoong.
Hi, Trung,
In gaan.py, self.md_x = tf.reduce_mean(self.f_recon - self.f_fake) According to Eq. (7) in your paper, maybe self.md_x = tf.reduce_mean(self.f_real - self.f_fake)
Is it correct? Thanks, Sungwoong.