Seeds, Weights, and Prompt Parameters

The Basics of Seeds, Weights, and Prompt Parameters in Prompts

‍Seeds, weights, and prompt parameters play a crucial role in prompting. Seeds ensure the consistency of generated images, while weights determine the importance of specific elements in a prompt.

‍Seed:

‍Definition: A numerical value that initializes the random number generator in an AI model.

Purpose: Ensures consistency and reproducibility of generated images.

‍Weights:

‍Definition: A value assigned to different elements in a prompt to indicate their importance.

Purpose: Helps the AI model prioritize certain elements during image generation. The ‘weight’ is more about how much it influences the final image. The ‘weight’ is more about how much it influences the final image;

Stop at :

The ‘stop at’ is when it stops influencing it. The "stop at" parameter sets at which iteration percentage to stop taking the image as prompt. For example, at 30 iterations and "stop at" value of 0.5, it will take the image prompt for the first 15 iterations (at the set weight, of course), then move on with just the prompt (and the image it has so far).

At 0 it's basically the AI doing whatever it wants, at max it's hyper detailed (HDR looking) monstrosity.

‍Prompt Parameters:

‍Those are parameters that you can add at the end of a prompt

‍--ar (aspect ratio): Controls the aspect ratio of the generated image. BlueWillow options currently are 1:1 (square), 2:3, and 3:2. Don’t forget to add a space between ar and 2:3 (e.g., –ar 2:3)

--model: Specifies the AI model version to use, such as "V2"

--no: negative.

--as-is: for non-English prompt optimization

The ‘weight’ is more about how much it influences the final image;

The ‘stop at’ is when it stops influencing it.