Detailed explanation of Midjourney's 4 latest parameter instructions

1. Describe the words in advance from the image The Describe command can help us extract Prompt (keyword) from an image and generate a similar image. The specific operating steps are as follows: I tried several different types of pictures and found that for most types of pictures, the prompt words extracted by describe cannot be accurately restored, but it is effective in identifying the subject, composition, material, style, special effects, etc. of the picture. Good. When we don't know how to use professional terms to describe a certain content in the screen, describe can provide effective keywords, and then we can refine the prompt words based on the keywords. In addition, it is best to upload images without text content to avoid affecting the generation of prompt words. When testing Van Gogh's classic oil painting "Sunflower", I also found a very interesting phenomenon, that is, there were no words related to "Van Gogh" or "Post-Impressionism" in the prompt words extracted using describe, but the generated image is very similar to the original picture in terms of composition and main content. I don't know if it's because the dataset has been specially marked. Use the describe function to extract the hints from Van Gogh's "Sunflower": The painting shows several sunflowers in a vase, in the style of golden age aesthetics, pop - culture - infused, light yellow and light maroon, grandiose color schemes, whistlerian, 8k resolution, picassoesque, vibrant color scheme --ar 2:3 --v 5 This painting shows several sunflowers in a vase, with a golden age aesthetic, an infusion of pop culture, light yellow and light chestnut, a magnificent color scheme, Whistler style, 8K resolution, Picasco-style, vibrant color scheme-ar2: 3--v5 2. Repeat When using Midjourney to generate images, one quadrants will be generated by default. The newly launched repeat function allows Bot to generate multiple quadrants to meet the needs of users who want to explore multiple generation effects at once. 3. NijiJourney v5 has a stronger anime style model Nijijourney is a model specially designed by Midjourney to generate anime images. Recently, a new V5 version-niji 5-was launched, which is more stylized than the default-niji model. 4. Permutations and combinations The Permutations function allows us to put multiple words or suffix parameters with the same nature in a set of prompts, and then the system will automatically arrange and combine these contents to generate multiple sets of corresponding images. This feature is currently only allowed to use it in fast mode by members with Basic and Pro subscriptions. It consumes extremely fast GPU minutes and requires attention when using it. When using the Permutations function, you need to pay attention to the following: ① Permutations can also be used in suffix parameters In addition to basic text prompts, Permutations can also be used in model versions, suffix parameters, and image prompts. ② Permutations can be applied to weight symbols When used as a weight symbol,{} can help us quickly figure out how much value can make the picture just create/eliminate a certain element. Example: A blue-skinned elf warrior leans against a tree, in the style of D8D fantasy art::, {blue-skinned::0.1, blue-skinned::0.3, blue-skinned::0.5} It can also be abbreviated as: A blue-skinned elf warrior leans against a tree, in the style of D8D fantasy art::, blue-skinned::{0.1, 0.3, 0.5} The final rendering effect is: A blue-skinned elf warrior leans against a tree, in the style of D8D fantasy art::, blue-skinned::0.1 A blue-skinned elf warrior leans against a tree, in the style of D8D fantasy art::, blue-skinned::0.3 A blue-skinned elf warrior leans against a tree, in the style of D8D fantasy art::, blue-skinned::0.5 ③ Permutations can be used multiple times Multiple {} can be used in conjunction within a set of prompts, such as: A {cryi