feel free to delve in to it and form your own conclusion. One observation is the amount of variation 'seems' related loosely to amount of tokens.
It doesn't. Noticed it long ago and some of us delved in to it ina post and VioneT explained it has to do with noise variations. The off part to me is the Perchance documentation saying a false thing, since that's highly unusual and jarring with the rest of it; hence mentioning it
/agree
very cool!
Though you may find Kotlin WebView does not play nice with Perchance AI. I tried to put a Perchance AI art gen in a kotlin webview long ago and it was detected as like 'adblocked' if i remember right.
nice experience you are gaining! Kotlin is very powerful. And cool idea with the improving organization. I probably wont look in to the Search since it's a bit out of my way, but kno that with Kotlin you can solve anything. Even if you have 30+ bugs, those are 30+ opportunities to learn what you did wrong and not do them again. also I highly suggest making a simplified Toast function you can place anywhere easy peasy for when solving issues (equivalent of console.log to tell yourself what variables are at each step so you can pinpoint when a value isnt getting from point A to point B).
i very often get vehicles driving coolly on water. it means they're fast; i think
lOl. now the other event ends on the 33rd. just for cuteness
I'm building something with this thank you
Cool idea!
topic: Unexpected Sequel
There is a style option and the background can be styled, but it seems to be styling a div 'behind' the gallery which is only visible in weird flashes when scrolling. try style = background-color:rgb(0,255,0); in a gallery to see what i'm talking about. Would love to make the background of beautiful people black to match the rest of it.
read this to improve your Anthro images
Looking back at VioneT's Guide on image prompting I came across where long ago i had come across the model underlying Anthro, Furry, Dryad, MyLittlePony and basically anything furryish, which has a very distinctive art style on Perchance. I had come across this peculiar post which explains that, instead of words like masterpiece, Pony Diffusion pictures were rated with words score_9, score_8, etc with score_9 being 'the best'. I see in Vionet's guide that this was quoted in so I doublechecked tonight now that I can seed correctly.
I believe we are using this which is why our anthro pics are so good.
here is an image starting with the word score_9
pretty good
now here is a score 4
notice it is more jagged and more like an amateur illustration but still decent.
But, and this I read from someone on civitai, the best results they were getting was to include not just score_9 but score_8_up and score_7_up because that made the group of pictures that were being drawn from big enough that the variation in some way made it 'healthier'. and here is when i had the, according to them, correct prompt for optimal Pony Diffusion pictures:
and i agree, this one does seem best.
so the prompt starts with: score_9, score_8_up, score_7_up,
and the antiprompt starts with: score_4, score_5
your welcome.
when i run for world president vote for me
enjoy :)
(only works with Pony Diffusion aka anything with the anthro style on perchance)
Dear Perchance, when an image is clicked there is a button in the top right to show seed. Can there please also be one to show GuidanceScale?
this is the most ancient, as far as i know uncompleted, quest i have found yet on Perchance. If completed, whoever completes it will receive good fortune and one miraculous event irl the following month.
https://perchance.org/beautiful-people
When I first entered Perchance I duplicated the t2i, then realized naming artists was important and, on my second generator got a list of 800+ artists and also got out of t2i so I could control things more finely. Then, to sift the 800+ artists for those with styles I like, I made Love and Light which, while clunky compared to Beautiful-People, does, if one gives dislikes and likes on the 4th and 5th images, eventually synthesize those 800+ artists in to just favs. Finally with 'artist combos' that look good, I started saving them to my own 'fav artistlist list' and, with that, made Beautiful-People.
And people do look good.
Also spiritually I understand Bhakti, where a believer focuses on Lakshmi (or whoever) to become them. Or in Christianity how one would focus on Jesus to become Christlike. Similarly with images of people. So I inherently don't sexualize people and I only give them positive characteristics. I basically surround myself with good romodels. Beautiful-People reflects this.
Also I initially started a version of Love and Light to collect Landscape artists but did not get very far in. So Beautiful-People is specialized in People and for landscapes is nonideal.
After the huge complexity of my previous generators, I wanted something that would FEEL good and be basically 'click button get beautiful images' so it is centered around this concept.
Eventually I upgraded my plugin for the prompt box with tons of libraries with customizeable names and colors. I also believe the best tools are fully customizeable by the wearer and may be made to be fully aligned with their own unique energy.
If you want to see a huge mass of interweaving perchancescript for prompts, touch the edit button. Also alot is built on my own plugins and objects. I have toyed with the idea of making a nice duplicateable framework like t2i. But Beautiful-People is not that. It is duplicateable, but I made it powerfully instead of to be easily built on by others. The artistcombos also are 'my fav artists' in the ratios i like because of how Love and Light works (and more artistcombos, which are more and more tuned, enter daily). This means the images Beautiful-People generates a year from now will be higher quality than the ones generated today. I believe in Growth, Scaling, and that learning to play the strongest lategame champion is the best strategy because then one only needs to learn how to survive to lategame and they become one of the top players in the game. So generators structured to scale up over time are basically where it's at, according to me, allo.
have a great day all :)
next things coming on Beautiful-People:
things i want or need:
edit: now with Meow button
The Meow button shows as an M in the lower right of generated images. This takes you to the Meow screen where you can mess with this specific picture. Most recently I got this nice but doubleeared pony.
so i touched the M button, generated, and got other versions of her that were better.
Now with this M button, all those pictures that just have something wrong with them even tho they have a great part are salvageable. Hard to overstate how cool this is.
I also like this pony but she has two horns.
so i hit the m button and typed 'one horn' at the end of the prompt
she is even facing a different direction now, but notice how it is the same pony with one horn.
4/8/24 now with customizeable prompt buttons
these are lolzy. was getting longneck people so
now my antiprompt literally includes: "(long neck), ((extralong neck)), very long neck, " and still got this image
maybe wrapping something in the antiprompt in () makes it happen instead of not happen? she clearly has an ((extralong neck))
https://perchance.org/linkbutton
To place buttons that are links in the corners of your pages :)
i personally am using this to help people find https://lemmy.world/c/casual_perchance and build a thriving, vibrant community :)
3Tho I do not think this is one of those areas.
Long ago I actually made a generator that was based off this concept of same seed = same image. After a decently long build time, it turned out that same seed does NOT equal same image and was an unforseen wrecking of that design concept; not just unforseen, but I had read this same sentence I am highlighting now and it is specifically this sentence that made it sound good to go but as an unexpected failure with wasted effort.
It can be explained to newbies without causing harm. But right now it isn't 'ommitting nonessential info'. It is portraying something incorrectly and, if anyone uses the info for a generator using the concept of Same Seed = Same Image, they get pwned by the statement having been unexpectedly false.
I do not know anywhere else in Perchance where something stated as true is actually false and if you use that info assuming it's true you get pwned.