Question about your implementation detail

[Masaki Baba]

Hi, I am recently interested in texture generation, and I found your paper, and found pretty good.
I have a few questions about your implementation.
I am not quite familiar with theano, but it seems to be different with your paper and following codes.

psgan/psgan.py

Lines 58 to 65 in bab3e57

 if config.nz_periodic > 0: 
     for i,pixel in zip(range(1,config.nz_periodic+1),np.linspace(30,130,config.nz_periodic)): 
         band =  np.pi*(0.5*i / float(config.nz_periodic) +0.5   )##initial values for numerical stability             
         ##just horizontal and vertical coordinate indices 
         for h in range(zx): 
             Z[:, -i*2,:, h] = h * band  
         for w in range(zx): 
             Z[:, -i * 2 + 1, w] = w * band  

Does it using the periodic noise part in your code?
And also when zx_quilt which is in your code is "not" None, how should I calculate periodic noise?
If I am taking wrong with your paper, please just let me know.

Thank you!

[Masaki Baba]

Actually, I just figure it out.
So please just ignore this issue.

thank you.