So there are a lot of examples for DCGANs for generating pictures and other 2-dimensoinal things, but there are just a few for generating time-series. I’m not super experienced with pytorch and GAN in general, but I understand all their basics. At the moment, I want to build a DCGAN which can eventually generate sinusoidal waves (because this looks like the easiest example of mimic a general time-series).
I looked at the general dcgan example and tried to make a similair dcgan with 1d convolutional layers in order to train them on sinusoidal waves. But unfortunately I didn’t succeed yet in making this GAN.
So I was wondering if there are any clear notebooks in which they are making a GAN to generate a simple time-series like a sinus wave?