The are many GAN related or DeepDream related opensource libraries. Since my background is in computer science and deep learning, I feel very comfortable to change these source codes myself. But CS background is good enough for what you may want. So if you have some basic computer science background, working with the opensource library is not that hard. You just need a good GPU card on your computer if you want to train a new model. I found some out of the box software packages a long time ago. They do not have the exact flexibility as most artists want.
This area changes very fast, so the best way to find these packages or libraries is just to google it. That is most people do. But again, GAN and DeepDream are two different approaches. Just want to point it out also.