ChatGPT, the OpenAI chatbot, has been making waves in the tech community since its debut. Its capabilities are vast, from generating code snippets to detailing the steps for using specific tools, frameworks, and languages, offering best practices, creating resumes, and even writing articles.
But my curiosity has led me to ponder on how ChatGPT could enhance my personal passion – photography. In this article, I’ll be sharing how to utilize ChatGPT to craft expert prompts for image AI generators such as DALL-E, Midjourney, Nightcafe, and DreamStudio AI.
Small info about the AI art Generators
In a nutshell, an AI Art Generator is a system that can produce images, videos, or music based on a given text prompt. It works by being trained on image-text pairs. Recently, I’ve had the chance to experiment with a few different AI Generators, namely DALL-E, Midjourney, Nightcafe, and DreamStudio AI. They were easy to use and the generated images were of great quality.
Let me tell you a bit about DALL-E2. It’s a text-to-image AI art Generator that utilizes the diffusion and Contrastive Language-Image Pre-training (CLIP) model. In comparison to its predecessor DALL-E (which is a version of Generative Pre-trained Transformer), it has a smaller number of parameters – 3.5 billion as opposed to 12 billion. However, it’s important to note that the prompt length for DALL-E2 is limited to 400 characters and the output images are very realistic. So, if you’re looking for more artistic images, DALL-E2 might not be the best choice.
Next, let’s talk about Midjourney. It’s an AI art generator that uses text-to-image techniques based on Generative Adversarial Networks (GANs). The communication between the user and the internal AI servers is done through a Discord bot. The images generated by Midjourney have a unique and striking artistic style. To be honest, I’m quite a fan.
Moving on to Nightcafe AI Generator, it used to employ neural-style transfer models but with the recent updates it uses CLIP + Vector Quantized Generative Adversarial Network (VQGAN) and Stable Diffusion models. One of the main benefits of using Nightcafe is that you have the permission to use the generated images however you please.
Lastly, DreamStudio AI – Stable Diffusion 2.0 is a text-to-image AI art generator from StabilityAI that is based on the Stable Diffusion model. In terms of quality, it’s equivalent to DALL-E 2 but it’s more artistic.
To put it simply, a prompt is a set of instructions, written in natural language, that serves as input for an AI generator. It’s crucial to provide a well-crafted prompt with the appropriate keywords in order to produce high-quality content using AI art generators. The output you receive will highly depend on the quality of the prompt you provide.
A good prompt should include the following elements:
- Subject: Represented by nouns, this suggests to the AI system what scene to generate.
- Description: This implies additional information related to the subject, such as adjectives (stunning, lovely), background description, etc.
- Style: Indicates the theme of the image. It can include artist names (Picasso, etc.) or custom styles such as fantasy, detailed, modern, contemporary, etc.
- Graphics: This stands for the computer graphics engine type that enforces the effectiveness of the image. Examples of such keywords include Unreal Engine, 3D rendering, and Octane Render.
- Quality: Indicates the quality of the image such as 4K, 8K, 12k or UHD. You can add image ratio size also in end
Create AI-Powered Art Generator Prompts with ChatGPT
Im lazy so using only midjourny , NightCafe 🙂
Prompt: cute baby robot smiling , covered by snow, modern, unreal engine, 12K
I Ask chat Gpt to make prompt lets see the result
Now we got the prompt and time to try it
Prompt: A friendly and curious baby robot, with a humanoid form, exploring a bustling futuristic city covered in snow. The robot has a happy and playful expression, with snowflakes gently falling on its metallic body. The scene is captured from a low angle, highlighting the robot's small size and vulnerability in contrast to the towering skyscrapers in the background , 12k , --v 4
you can add –v 4 in midjourny for better result from version 4
I got poor result from NightCafe
well image is better than before in midjourny haha thank you chatgpt
I like midjourney comment down your favorite tool
Now it’s your turn to take the reins. Don’t be afraid to go above and beyond by elaborating on the description, adding new styles, and so on. The beauty of it all is that the limit is set by your own imagination. Don’t be afraid to let it run wild.
I hope you’ve found this article informative and enjoyable
Comments 1