This version brings further improvements to detail levels, better realism support, better Sci-Fi, and fewer quirks overall.
Tips:
Keep CFG fairly low (e.g. 9 and lower, ideally around 7).
If your results don't look very clean or have an artifact, you'll probably need to re-run with more steps -- this model can be more demanding than others.
You do not need to prompt a bunch of detail/quality phrases, or use 'negative embeddings' or massive negative prompts. You can be short and simple. It'll do just fine without handholding, you can trust me.
To control detail level for less detail-dense art styles, try putting 'photo' and 'realistic' in your negative prompt.
If your images are coming out a little dark or contrasted for your liking, try putting 'low light' in your negative prompt, or 'dark' for brighter colors if it's not a lighting issue.
May occasionally need to put 'paint' in negative prompt when prompting for paintings and you're getting weird extra colors or actual paintings in the scenery.
Oddities:
Eyes are often a bit off and in need of correction. Dunno how to deal with that, just happened when incorporating photorealism models.
Trying for realism with natural landscapes doesn't normally look great.
The model can be a bit demanding for step counts.
The mix has been redone from the ground up, introducing better semi-realism capabilities when prompted, greater level of detail by default, and while fixing a few oddities (now you can type 'cyberpunk' without it automatically meaning a million mini signs being put on everything, yay!). I'm also switching gears to focus more on using only the finest community-sourced, organic, cage-free, grass-finished 1.5 finetunes. It's easier to control and keep track of model behavior that way.
