GANDALF

This is an artwork by the artificial intelligence GANDALF. GANDALF is a deep learning network with a GAN architecture (based on “Generative Adversarial Networks” arXiv:1406.2661) and has been trained to produce its own digital artworks since September 2020. I am collaborating with GANDALF as its trainer. To train GANDALF, I create various image dataset inputs. Based on those pictures GANDALF develops its own artistic visions. With every session I made some little changes (concerning the learning rate, workes, input data or input folder structure for example) and monitored closely how each change affected the generated output. After every session I made a selection of generated pictures and introduced them into the input data.
Hannah Haifisch

hasn't created any creations.