🤖 I've been exploring various AI art tools such as Dall-E and Mid-journey. After extensive experimentation and testing, I find myself drawn towards Stable Diffusion as I can guide and steer the AI's direction as if I was a teacher by using a combination of descriptive prompts and negative prompts.
Negative Prompts Vs Prompts 💡
✨ Negative prompting is just as crucial as descriptive Prompts in avoiding common errors in AI art, such as lifeless eyes, multiple limbs, distorted hands, and extra fingers. When crafting your Prompts, aim to provide as much detail as possible about the image you want to create, while still giving the AI some artistic freedom to imagine. Think of yourself as an art teacher, nurturing your student's imagination.
✨ Make sure to always include a [Hero element + mood/feeling + setting + tone/style + reference/inspiration] in your prompts. For composition, I have been using a combination of loose sketches, photos of backgrounds, friends and simple, basic 3D objects posed in Cinema 4D (or other 3D tools) to provide guidance for the AI. Additionally, I employ advanced techniques like texture-to-texture to refine my images.
⬇️ - Please see all my examples below - ⬇️
Theme: "Portrait of a woman in a Moroccan tent" using Stable Diffusion v1.5 🤖🎨
Key Prompts: Portrait of a woman, kneeling, Moroccan tent, war-torn, dirty face, wild hair, incredible detail, highly detailed, Realistic.
Key Prompts: Portrait of a woman, kneeling, Moroccan tent, war-torn, dirty face, wild hair, incredible detail, highly detailed, Realistic.
Theme: A samurai in a Cyberpunk futuristic city in neon-soaked lighted streets using ControlNet and Stable Diffusion v1.5 🤖🎨
Using: ControlNet, Image 2 Image
Key Prompts: Samurai Portrait, Cyberpunk, neon lighting, realistic, wet, dense atmospheric
Using: ControlNet, Image 2 Image
Key Prompts: Samurai Portrait, Cyberpunk, neon lighting, realistic, wet, dense atmospheric
Theme: I'm a super fan of The Mandalorian , Star Wars and the Sci-fi genre in general so I had to do some experiments, and have fun testing Prompts using Stable Diffusion v1.5 🌌🐸👶
Theme: Turning my Digital illustrations into real people using Stable Diffusion v1.5 🤖🎨
Using: Image 2 Image
Key Prompts: Extremely Detailed, Instaport, Realistic, Portrait, FujiFilm, Raw Photo, Masterpiece
Using: Image 2 Image
Key Prompts: Extremely Detailed, Instaport, Realistic, Portrait, FujiFilm, Raw Photo, Masterpiece
⬇️ - Move the Slider - ⬇️
📝 Note all images (Besides reference images for ControlNet) presented here are created by me
and all people depicted are AI-generated and do not exist! 🤖
Thanks for Viewing
Thanks for Viewing
