I have a project consisting of generating high quality free ebook covers out of its content. On my 16GB of ram machine with no gpu, i have tested the opensourced stable diffusion models without any success. All return bad quality covers with blurred faces and scenes that do not match the prompt what
