You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Why are two FFTs used in the model? Are there any differences in their output results? Why do we need to input the decoder together?
Which package and line is FFT in the code file?
The text was updated successfully, but these errors were encountered:
The model outputs the z0 and noise simultaneously, so there are two FFTs. The FFT module is here
Hello!
I have a question to consult. If we don't add noise to the image, would the FFT module be able to filter out non-edge parts ? As edges are high frequency parts of images, I am guessing this FFT module would learn to filter out those low frequency parts to get edges.
Why are two FFTs used in the model? Are there any differences in their output results? Why do we need to input the decoder together?
Which package and line is FFT in the code file?
The text was updated successfully, but these errors were encountered: