Midjourney has announced an early test version of Model Personalization, a feature that aims to make the image generation process to each user’s unique preferences. Every time you write a prompt, it- often remains ‘unspoken,’ and so with new features, Midjourney will try to find a solution to these problems.
To enable this feature, type –p after your prompt or turn on personalization for all prompts under /settings. When personalization is enabled, a unique ‘code’ is added after your prompts, allowing you to share the personalization effect with others. According to the announcement, the strength of this effect can be adjusted using –s, with values ranging from 0 (off) to 1000 (maximum), and the default set at 100.
Also, Midjourney has added more advanced options for –sref (style references) and –p (model personalization). Users can now combine multiple style reference codes, as well as blend style reference image URLs with random codes.
Also, you can combine multiple –sref codes into a single command, such as –sref 123 456, allowing for more complex style references. Additionally, you can blend style reference image URLs with sref random codes, for example, –sref 123 url, to achieve unique visual outcomes.
According to the announcement, further algorithm updates are expected over the coming weeks.
