Prompt: an art-deco style poster of a smiling lemming with the text "Welcome to Lemmy!" Below it. --ar 3:5 --v 6.0
I’ve been waiting eagerly for this update, looking in envy at those bing and chat-gpt images! :)
Not so much for the text, but for the greater coherence in the image and better understanding of the prompt.
Apparently the prompting is completely different and you’re expected to use more natural language. I’ll be happily experimenting for the coming days :)
Edit:
A few more examples. From what I can see the --stylize parameter has quite a bit of influence on how accurate and integrated with the image the text will be.
--stylize 50
--stylize 200
--stylize 350
--stylize 750
At really high levels the text is almost always accurate, but contained to its own little text box. Also it ruins the art style I intended to go for and it looks more like Fievel than a lemming :)
I know, I know! :)
But I really love the style Midjourney applies to all of its images, in general I find them prettier.
Also I like the control I have with aspect ratios, remixes and image input.