Delve into the intricacies of prompting and prompt expressions on Scenario
If you haven't learned the basics of Prompting, start here: Write a Good Prompt on SD 1.5 and Writing a Good Prompt for SDXL and LoRA Models
Understanding Prompts
- Case Sensitivity
While prompts in Scenario aren't traditionally case-sensitive, each word or phrase in a prompt corresponds to a 'token'. Different cases might yield varying outputs, but differences might be subtle. Familiarity with prompt engineering is advised for optimal results.
- Negative Prompts
These prompts indicate terms that shouldn't appear in the generated image.
- Sampling Steps
This refers to the AI's steps in crafting an image. The required number varies based on image size and complexity.
- Reference Image Utility
The Reference Image feature (Img2img) lets you use your image as a style and composition guide for the output.
- Consistent Image Outputs
By default, images generated are unique. To narrow output variety, use a Reference Image (img2img) with a high Influence setting, Composition Control Modes, and Seeds.