Prompt Guide
What is a prompt?
A prompt is a user input provided to communicate with AI models like mnml.ai, designed to describe the desired outputs.
Think of it as a guiding instruction that helps the model understand the task or concept you want it to focus on.
Prompt Examples
Let's take a look at an example prompt to illustrate its structure and components.
Good Prompt
“villa, modern architecture, golden hour, stone facade, glazing”
Poor Prompt
“Please design the villa for me”
“I want a beautiful design”
“As a designer, i need the villa design to follow the concept that emphasizes a minimal and cozy aesthetic, encompassing passive voice to present a clear and concise architectural vision.”
How to write good prompts?
When crafting prompts for AI models, follow these guidelines to maximize effectiveness:
Start with simple keywords
Use precise language, defining the building type, architecture style, scene context and mood.
Example: admin building, parametric design, downtown area, daylight.
Be Clear and Specific
Clearly communicate the desired output or task to the model. Don't be verbose, soemtimes long prompts returns poor outcomes.
Use Negative Prompts
Sometimes negative prompts are essential combined with your description in order to remove elements from the photo.
Prompt: villa facade, parametric design, curves.
Negative prompt: straight lines, rigid form.
Iterate and Refine
Don't hesitate to refine and experiment with your prompts. Iterate based on the initial results, tweaking them for better outcomes. Continuously learning and adapting your prompts can improve the model's outcomes.
Understand Model Limitations
Familiarize yourself with the capabilities and limitations of the specific AI model you're working with. This knowledge will help you tailor your prompts effectively and achieve better results.
Pro Tip
Remember, writing good prompts is an iterative process. Experimentation will help you to generate impressive and meaningful outputs.