z-image: adjust strenght to 0.8 - 1.
this was requested, my default style loras was trained in 500 steps, but this goes beyond 1000.
no need to trigger, just prompt and go.
yeah you can mix with your model, for merge guys.
Description
FAQ
Comments (4)
can you make qwen lora for Bryce3d style? it's 90's 3d modeling program that produced very distinctive artstyle.
and also maybe 90's-00s y2k cgi style in general, not in-game graphics but from FMVs and renders of that era
How do you get good results in 500 steps? Any LR technique or adjustment? What is the dataset size?
you can see in my recent article on my profile a guide, this article covers 150 steps
at mid strengths (0.3 to 0.6), this Lora actually pretty decent at softening the image and adding more personality. Going too high will make it too noisy.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.