In the past month, in addition to bringing us the new Niji V6, Midjourney V6 and the official website have also been updated with some new features and commands, the most important of which is the new style reference parameters.
Adding "--sref" and a link to the reference image after the prompt will cause Midjourney to use the reference image as a "style reference" and try to produce a work that "matches" its style or aesthetics.
This feature is especially useful when you see a favorite that you really like and want to try to generate a similar or identical style, but you don't know what that style is.
For example:
Another example:
This parameter is only available for Midjourney V6 and Niji V6 versions.
sref".
Multiple links can be used to have midjourneys referenced at the same time. The parameter is --sref urla urlb urlc
You can also set the weight of each graph, such as "--sref urla::2 urlb::3 urlc::5".
For example:
The strength of the overall stylization can be set via --sw 100 (default is 100, 0 is off, and 1000 is maximum).
Tip: If your prompt leans towards photorealism and you want a conflicting style, such as an illustration style, you may still need to add some text to your prompt. The style reference mode supports pads, but has no effect on image-only prompt generation, and only affects tasks that contain at least one text prompt.
v6 now supports panning the zoom change (area), and after zooming in on the ** generated by v6, you will be able to see the corresponding button below the **.
Panning is now easier to use and provides higher quality results (with more coherence and fewer repetitions).
The "Feedback" feature allows you to submit items that you want Midjourney to prioritize.
The Midjourney team uses this data to craft a large list to help their engineers determine what to focus on.
The team also publicly reviews the list during each office hour and presents it with cool visualizations.
The alpha version (with image creation) is now available for the 5000 Club. The 5000 Club includes everyone who has produced at least 5000 images on Midjourney.
You can check how many images you have made with "info".
The search bar has been moved and is now always displayed at the top of the page.
More options to filter your images (under My Images).
Added viewing options to change how images are displayed (under My Images).
You can create folders and smart folders to organize your images (under My Images).
You can "click and drag" to select multiple images for bulk publishing** or bulk unpublish (under "My Images").
Midjourney also teased the upcoming release of the "Role Consistency" parameter "--cref". With this feature, the characters you generate can finally generate stories and storyboards in a coherent manner!
If you think this article is helpful to you, welcome to like, bookmark and share. In the meantime, please follow me for more updates and insights on artificial intelligence!
February** Dynamic Incentive Program