. Today, let's take a look at V6.
This version is Midjourney's third model that has been trained from scratch entirely on an AI supercomputing cluster. The project took nine months and is now finally meeting with fans.
Official Blog - https://mid-journey.ai/midjourney-v6-release/
In Discord settings, you can now choose to try the V6 Alpha version.
The main updates brought by V6 include:
: Instructions provided by users will be understood and executed more accurately. ：The output images will be more consistent in style and quality. ：The quality of image generation and blending has been improved. Click Remix mode in Setting change to Check the final effect: a combination of real-life scenes and anime effects The text on the generated images is no longer garbled. : Maintain better clarity and quality when enlarging the generated image.
In addition, the following functions/parameters were supported at the time of the V6 release:
--ar--chaos--weird--tile--stylize--style rawVary (subtle)Vary (strong)Remix/blend(Currently visible only in the V5 version)
The functions planned to be supported in the next month include:
PanZoomVary (region)/tune(V6 version next month will also have these features)
It is worth noting that the usage of V6's Prompt has changed from before, and users may need to relearn and adapt:
There is no longer a need to add unrelated words such as "award-winning, photorealistic, 4k, 8k". The Prompt needs to be more specific and clearly express your requirements. parameter. Lower values of the parameter (default is 100) may better understand the prompt, while higher values (up to 1000) may perform better aesthetically.
Let's take a look at various effect images provided by users on the X platform together:
