GlassTech - World Morph 💎
This is my second attempt at creating a world lora following the "Konyconi-LoRA" method, if you don't know Konyconi you should really check out his LoRAs they're simply mindblowing 🤯
About : this LoRA "GlassTech" is meant to influence or change the entire course of the final image to a dimension that relies solely on: glass ,glass circuitry, synthetic glass, glass fiber ...
with that in mind you must use the trigger word glasstech then pick one of two directions:
Sci fi, by adding
scifito the prompt witch was my primary intent with LoRA.Fantasy, by adding
magicorfantasythe prompt witch mostly works well with architecture , armor or creatures giving them an arcane eleven look.
further more, you can include the one of the following decorators to influence the style even more : transparent iridescent gold wires
Usage Tips :
use weight of 0.7 to 0.9 to influence the image .
use weight of 1 to 1.5 to use as world morph .
adding
transparentwithscifiworks rly welluse hires fix for large detailed scenes
Mostly use with "Eular a" with a step count over 20
rly shines at img2img at weights over 1
Samples workflow :
add the LoRA at weight of 0.9 to the prompt along with
glasstechscifitransparentset the sampler "Eular a" , steps 30 , CFG Scale 5.5
enabled hires fix at 0.55 , steps 20 , Upscale by 1.5
Generate, then send to img2img
crank up the LoRA weight to 1.2 or 1.4, steps 35 , Denoising strength 0.42
increase the resolution for example to 1024x1024 , then Generate
send to extras for further upscaling if needed
have fun and consider rating ⭐
Description
FAQ
Comments (9)
beautiful.
also.. thank you for the shout-out.
Your LoRA method is absolutely brilliant, thank you for sharing it ❤
Generates cool environments for NSFW portraits. Love it 💚
Koniconi is replicating 🥲😁
man konyconi is everywhere now 🤣
Thanks for the lora, great job. What models were those samples created in? Same as some of konyconi's such as revAnimated_v122?
yes I also used RevAnimated, it does wonders in generating small details.
Could you make a newer version after removing the Bing watermarks from your data set?
Why not just compile everything? i feel like if i want to make something futuristic, having 12 loras is probably not efficient.
i have no problem running a bigger size lora
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.


















