golden radish
This prompt provides detailed guidance for generating prompts that specify various characteristics of the images to be created by Midjourney's image generator. The key areas of focus include:
The understanding of Midjourney's various parameters such as --chaos, --ar, --s, --weird, and --tile and how they influence the generated images. Each parameter provides a different layer of control over the resulting image, from the level of randomness to the aspect ratio and stylistic application.
The concept of multi-prompts is also important. It allows ChatGPT to consider multiple concepts separately in a single prompt, using weights to balance the prominence of each concept in the generated image. This gives the user fine-tuned control over the resulting image's focus.
The prompt instructs the AI to generate five prompts for each keyword or phrase provided, three of which should relate to specific art styles. The remaining two prompts will be regular ones related to the user's keywords or phrase.
The training prompt also provides guidelines on the formatting and structure of the generated prompts to ensure consistency and easy integration with Midjourney within Discord. It will output the 5 prompts in python markdown and start each prompt with " /imagine prompt: " so all you have to do is copy and paste into Midjourney
It is instructed to only output various aspect ratios for each prompt, but it has the knowledge of all the parameters. So if you would like to see high stylization, just tell it you want style added and so on.
This works best with GPT-4, as GPT-3.5 more than likely will not remember the instructions and/or just not follow them the way they are written.
Please feel free to let me know if you have any questions and if people like it, I will do my best to keep it updated as Midjourney progresses.
plain blade
upper slate