first image with lora second image without lora
vivid primary colors, asymmetric, cinematic, movie scene, angled shot of stylish goblin math teacher standing in front of a blackboard, wearing sunglasses, in a magic made of strawberries wonderland, gold highlights, cute, detailed light, cinematic lighting, hard shadows, haze, hazy atmosphere, complex reflections, precision, colorful, digital illustration, photo manipulation, altered realities, fantastical scenes, digital artistry, striking visuals, sharp focus, detailed background, light reflection, masterpiece, best quality, absurdres, RetroBG, strawberry-themed elements
first image with lora second image without lora
ral-cyclops, score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, solo,looking_at_viewer,full_body,yellow_eyes,flower,outdoors,horns,sky,cloud,tree,pokemon_(creature),no_humans,fangs,cloudy_sky,grass,sunset,twilight,tusks <lora:ral-cyclops-pony:1>
;;;;;;;;;;;;;;; updated details below july 5 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
A very interesting discovery. This lora's strength in different categorizes will fluctuate based on its need to balance the other lora's when needed to fulfill a request.
So if you are using this lora with other lora's that have a wide range and are very capable on their own, then very little of this model's strength is needed for a dramatic impact. i am finding that with strong mathematical lora's and wide range lora's that have a complex core already. i can set this down to .05 and see it bring the scene together nicely ( this lora is very good at being a middle man between elements of many loras )
still in an advanced stage of understanding the full potential of this lora, basically finding out that this lora is extremely versatile and self adjusting to how it will interact based on the other models connected to it.
example link : going into high values with wide range lora's can yield some amazing surprising elements like opening pandora's box just for a second and something new and interesting happens to the scene that fits the prompt in a new configuration
from .0 to .9 depending on what its connected to you will find there are usually 4 solid scenes hidden incremental, so shifting to much will bring you into a slightly different configuration
the first thing this lora does is look at the scene and check if any connected loras are bleeding into each other when not desired and fixes that while checking all categories below for were it can be useful and fixing/correcting/removing if necessary
ex: jpeg artifact: i have seen it clean those up nicely
will keep adding info as i go along : )
with lora :: anime style, smiling princess wearing an absolutely gorgeous dress standing on an outcrop overlooking moon-bear island :: link to image
without lora :: anime style, smiling princess wearing an absolutely gorgeous dress standing on an outcrop overlooking moon-bear island
``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````
imagination -- new name --simple- what its doing -- countless hours analyzing data-- update :
setting to between .20 - .30 for all below at optimal levels of imagination
keep in mind that this is compatible 99.9 % of loras and stacks (still use your favorites)
will add all variations to list.... sd, pony, flux, etc...
setting above .30 imagination and the imagination will really begin to affect every aspect of the typical image categories below:
-- detailer
-- skin detailer
-- additional prompt adherence thru imagination
-- eye detailer, -- fix eyes
-- correct all anatomy in general
-- its core is creative based -- art style enhancer
-- think of it this way.. what couldn't use a little more imagination
-- it knows how to interact with all elements on the stage(scene) using imagination
-- hands -- in the a to c grade sometimes an f if you start to get really creative
-- text ... like from an B grade to consist C with couple A ...so improvement
-- fix guns, fix eyes, fix anatomy, makeup anatomy (you wanna create something new), create and finish scenes from a point on the scene
-- hey boss where do you want me to put the couch? use your imagination : )
#############
After much testing was able to grasp enough to take testing to the online gen and i pinned examples below : )
lora is made from artwork pre computer, and then still all oil, watercolor, colored pencil, markers, crayons from olive garden from the floors that i picked up as a bus boy.. etc...
also i had a kinda i almost finished it kinda style..lol i get to it.. style...hahaha
!!! update maybe other knew.. dunno but interesting find, depending on what other artist i want to mix with depends on what parts of the works i gave it that it uses
--example what lead to this- mc escher is a mainly black and white artist and when i used his name it dipped into my ink work which had similar complexity