In this project, I created a Generative Adversarial Network (GAN) to generate abstract art pieces. The image data used for training the GAN was collected from WikiArts.org, which provided a diverse collection of modern art encompassing various styles. I specifically chose the images from the following categories: abstract expressionism, action painting, minimalism, analytical cubism, color field painting, and cubism. To ensure consistency, the images were resized to 1024x1024 pixels.