Midjourney update launches Style Tuner to create custom styles
Midjourney has launched a new feature called “Style Tuner”. This feature allows users to create their own styles and apply them to new image generation.
Style Tuner affects various aspects of the resulting image, such as color and character detail. Users can adjust Style Tuner to adapt to different image generation needs (such as image proportions, specific prompt words, etc.).
Using Style Tuner
To use Style Tuner, type "/tune" and then enter the prompt. You can then select the number of base styles to generate, and after confirming your selection, the estimated GPU time before job execution is displayed.
Once completed, a custom Style Tuner web page is created and its URL is sent to the user via direct message. On the Style Tuner page, users can choose their favorite style to create their own personalized style. For each style, there is a style code on the web page.
The user can then apply the selected style code to the new image by typing "/imagine" followed by the style code, for example "/imagine frog --style CODE".
The following example shows how to apply the generated style code to a central theme (a frog in the style of a children's book drawing).
According to Midjourney, styles may only be optimized for certain cues and may not always translate to other cues as expected. For example, a style optimized for cat images may not work well for city images. However, Midjourney said it should work well for images of dogs.
In addition to the main functionality of the style adjuster, the command "--style random" can be used to create random style code without using the style adjuster on the web page. Multiple style codes can be combined using the command "--style code1-code2". The strength of the style code can be controlled with the command "--stylize".