>>16707485

My CPU takes hours to train really basic models. Using GPU acceleration it us substantially faster but still limited. My interest in DCGANs was a project to generate high resolution images of galaxies from space telescope astronomical data.

I wish I had access to a supercomputer or one of these trillion dollar GPU datacenters. The computing power required for this is massive, matrix calculus in the thousands of dimensions. My desktop PC simply does not have the processing power.