How to write a prompt correctly

2024-03-22

To get the desired results, it is important to compose the textual prompt correctly. Basic principles include understanding tokens, using commas correctly, and creating accurate and detailed descriptions. For best results in image generation, follow these principles:

1. Use tokens, separating them with commas.

Tokens are units of text into which the AI breaks down the text you enter. They can be single words, parts of words, punctuation marks. Understanding tokens is important because the model generates images based on each token. Commas play an important role in generating a hint because they help structure the information and make it more understandable to the model. In prompts, commas are used to separate descriptive elements and to indicate additional characteristics. For example: girl, reading a book, in the library.

2. Use a detailed description, specificity.

Specify the specific details you want to see in the image. The more detailed the prompt, the more accurate the result will be.

For example: beautiful, ten year old girl, 18th century clothing, sitting on the floor, reading a book, in the library, night.

3. Use adjectives.

Adjectives help to clarify the characteristics of objects. Use them to describe colors, shapes, sizes, and other qualities.

For example: beautiful, ten year old girl, long blond hair, 18th century clothing, long blue dress, sitting on the floor, reading a book, in the old library, night

4. Specify a style or mood.

If you have a style or mood preference, indicate it in the prompt.

For example: beautiful, ten year old girl, sitting on the floor, reading a book, in the old library, Fairy Tale, night

5. Use keywords.

Include keywords that clearly describe the main content of the image.

For example: beautiful, baby girl, reading a book, in the old library, Fairy Tale, night, magic

By following these principles, you will be able to create effective and accurate text cues for image generation.