|
|
 Like, zoiks! |
#
?
Nov 17, 2021 08:20
|
|
|
|
| # ? May 24, 2024 01:07 |
|
|
|
Finger Prince fucked around with this message at 08:23 on Nov 17, 2021 |
|
#
?
Nov 17, 2021 08:20
|
|
|
|
|
|
#
?
Nov 17, 2021 08:26
|
|
|
|
|
|
#
?
Nov 17, 2021 08:36
|
|
|
|
   
deep dish peat moss fucked around with this message at 09:03 on Nov 17, 2021 |
|
#
?
Nov 17, 2021 08:57
|
|
|
 A life in three panels. |
|
#
?
Nov 17, 2021 10:59
|
|
|
|
|
|
#
?
Nov 17, 2021 14:17
|
|
|
 
|
|
#
?
Nov 17, 2021 14:23
|
|
|
 
|
|
#
?
Nov 17, 2021 14:34
|
|
|
|
|
#
?
Nov 17, 2021 14:35
|
|
|
|
|
#
?
Nov 17, 2021 14:38
|
|
|
|
I have a bunch of weird ones but I'm out of uploads |
|
#
?
Nov 17, 2021 14:55
|
|
|
|
|
|
#
?
Nov 17, 2021 14:57
|
|
|
Escape From Noise posted:I have a bunch of weird ones but I'm out of uploads Upload to imgur you can host as much stuff as you want there. Just be sure to right click copy image link, and if you copied one that starts blob: copy it again you copied it before it was finished uploading (imgur does some caching to make it seem faster but if you are quick you get bad data). 
|
|
#
?
Nov 17, 2021 15:08
|
|
|
|
my method is to use windows' screen snip tool (windows ket + shift + s) to select the area i want, which puts the snip into your clipboard, and then on postimg.cc you can just hit paste and it'll upload ezpzp (the last p is for pal)
|
|
#
?
Nov 17, 2021 15:13
|
|
|
|
|
|
#
?
Nov 17, 2021 16:29
|
|
|
|
I'm phone posting and they're all on my phone |
|
#
?
Nov 17, 2021 16:30
|
|
|
Escape From Noise posted:I'm phone posting and they're all on my phone imgur app will work, the site is very not good for phone posting
|
|
#
?
Nov 17, 2021 16:33
|
|
|
deep dish peat moss you are really, really good at these sometimes naming a specific artist leads to really good results 
---------------- |
|
|
#
?
Nov 17, 2021 16:45
|
|
|
|
|
#
?
Nov 17, 2021 17:01
|
|
|
|
|
#
?
Nov 17, 2021 17:09
|
|
|
|
|
#
?
Nov 17, 2021 17:09
|
|
|
|
|
#
?
Nov 17, 2021 17:22
|
|
|
|
|
#
?
Nov 17, 2021 17:23
|
|
|
|
|
#
?
Nov 17, 2021 17:25
|
|
|
|
|
#
?
Nov 17, 2021 17:31
|
|
|
|
|
|
#
?
Nov 17, 2021 17:35
|
|
|
help i can't stop
|
|
#
?
Nov 17, 2021 18:04
|
|
|
|
|
#
?
Nov 17, 2021 18:06
|
|
|
 gotta stop now
|
|
#
?
Nov 17, 2021 18:10
|
|
|
Dr. Honked posted:
That's Fred
|
|
#
?
Nov 17, 2021 18:12
|
|
|
|
|
#
?
Nov 17, 2021 18:14
|
|
|
|
 I mean, it's not wrong
|
|
#
?
Nov 17, 2021 18:37
|
|
|
pixaal posted:That's Fred his mouth says "scooby don't" but his eyes say "scooby do"
|
|
#
?
Nov 17, 2021 18:38
|
|
|
Manifisto posted:
I swear this is how I pictured Roland in The Dark Tower.
|
|
#
?
Nov 17, 2021 18:45
|
|
|
|
|
#
?
Nov 17, 2021 19:00
|
|
|
|
|
|
#
?
Nov 17, 2021 19:04
|
|
|
|
froward posted:deep dish peat moss you are really, really good at these  I've been messing around with AI image generation from text prompts for like a year now so I have a pretty good idea for how image generation AI parses textI don't know how to best explain it but AI often does better with more vague prompts and specifically with words that evoke visual imagery without having specific visual imagery tied to them. In other words AI is very good at evoking vague categorical feelings over specific looks or actions. It knows what people look like when they are "sad" or "embarrassed" or "shameful" or whatever - but for something more specific like uhh, the look of knowing you just let a fart out on the elevator? It just doesn't know what exactly that looks like. Like the posts earlier on this page with "homosexual donald trump" - since "homosexual" isn't something that is visually descript in a person's face the AI doesn't really know how to interpret it - "don't homosexual people have the same faces as heterosexual people???" it wonders. So it's just giving you a random chaotic face with some of Donald Trump's features, and when an image generation AI gets chaotic and breaks down like that you're more likely to get images that look weirdly AI-generated, like imperfect textures and stitching and parts stretched or whatever. But what it does know is "flamboyant" and through using that you can get a pretty consistent categorical look  So it all comes down to experimenting with different vague descriptive words to find the right adjective to get what you want out of it If you're bored try messing around with random adjectives before a name like "glam" or "glowed up" or "jonesing" or I mean literally anything, just poke at it until it surprises you then run with that one on some other faces  Okay, so Glowed Up is a bad choice because "glow" is a distinct visual descriptor that is not the same as "glowed up" and it seems to sort of get the two mixed up - you can tell it was kind of trying to give them glowing skin and it muddies up the whole picture  Glam gives us ritzy 60s-80s portraits wearing makeup:  Jonesing gives us... uhh, mostly thinner skeletal figures: 
deep dish peat moss fucked around with this message at 19:27 on Nov 17, 2021 |
|
#
?
Nov 17, 2021 19:16
|
|
|
|
I'm getting good material from my Burger Wars stuff |
|
#
?
Nov 17, 2021 19:34
|
|
|
|
| # ? May 24, 2024 01:07 |
|
|
 Mario was in the sun too long
|
|
#
?
Nov 17, 2021 20:10
|
|