betafaceapi.com
>Upload an image
>AI finds and lists stuff about faces in it
Beta face api thread
nice try fbi
Nice try, hackerman
got him
Based AI
> attractive: no 100%
my sides.
The AI can detect traps
> no nigger:100%
So I put in some pics of people That I wanted to fuck in high school and one of them was a full on trap.
Just why
You're gay, and that's OK
>big lips: no
I tested more images and it says female and then male again.
I am so confused
I can do that with ease. This AI is unmatched compared to my God-given ability.
>Post pic of me
>Results appear
>Attractive: Yes
What?!
>uploading is fail
No Thanks China
Any chance of there being an offline version of this?
>Doing free data collection for a meme company.
HOLY SHIT
did the aliens already got out of area 51?
we've been here all along, user
>attractive: yes (73%), bags under eyes: no (97%), bald: no (93%), bangs: no (20%), beard: no, big lips: no (17%), big nose: no, black hair: no, blond hair: yes, blurry: no (80%), brown hair: no, bushy eyebrows: no (85%), chubby: no, double chin: no, expression: neutral, gender: female (87%)
this wont be around for long.
>attractive: no (92%)
on suicide watch lmao
>big nose: no
>big lips: no
>black hair: no
uh...?
>agent carl, what is the best method to faces of these chan losers?
>that's easy agent smith, just make a quick program to change the age/gender of a face or give them metrics about faces. then they upload their own face.
>fucking brilliant agent carl
those things have been around for years now, back when i started transitioning a couple years ago i used to use these to gauge how well i pass, most trans people actually use these a lot
What do the %'s mean?
i've been on Jow Forums longer than you have, nu-Jow Forums, so has the entire /lgbt/ board, /cm/, /soc/ and Jow Forums were always as gay as it is today, i'm not part of your Jow Forumsthe_Donald conspiracy, but feel free, i'll be living in your mind RENT FREE
sorry you fell for ROGD
Confidence, so how "confident" the AI is in it's "decision", google AI Confidence if you're interested, it's fascinating if you're into maths algorithms and such
I've been feeling trans since way before i knew LGBT people even existed, so that ain't it chief
*yawn*
that's believable, some of them can make weird faces to pass in photos but 0% pass in person
throw enough shit at a wall,
eventually a small amount will stick
true, but it's also true that most trans people actually pass, they just never make a thing out of being trans because why the fuck would they want to, they got what they wanted, switched genders, dropped their statistical chance of suicide to near normie levels and went on with their lives, you likely have seen trans people and never known, but then again i don't imagine your kind goes outside your mom's basement much
nope, if you made a betting game out of it the odds would always be against the tranny
>it's fascinating if you're into maths algorithms and such
t. CS freshman
not even close to being true, i'm literally part of the LGBT society at our uni and i don't know a single trans person who doesn't pass, FtM and MtF both, and i don't just mean they pass AI tests or in photos, i mean, i hang out with them and i've never seen them get misgendered despite not really making an effort to pass off as particular gender, maybe in 56% land you're all part nigger so it's harder to pass but here in Europe i've never met an unpassable trans person
Final Year coming up actually, after my paid work internship work experience year, already on Dean's List for second year outstanding achievement
If you're not interested in math algorithms you don't belong on Jow Forums
another day, CIA
this thing is bullshit
>Attractive: No
Even AI hates niggers.
most males are unattractive
Most 3d people are unattractive.
no, just most males
the attractiveness shit doesnt work
seething uggo
dilate
>here in Europe
?
hmm...
betafaceapi has a few issues from my experience.
It will analyze faces were there aren't any.
It misidentifies race quite often. Specifically, blacks as asians.
I don't understand why the binary answers don't just have one confidence for yes.
Based and cunnypilled
Based AI
it's right tho, the dude looks really wonky
It might've been trained weirdly so it has spectrums of, for example, not attractive to attractive and not non-attractive to not-attractive and outputs the one with bigger values, or it says values of [50;100]% are [0;100]% attractive and values of [0;50)% are [0;100)% non-attractive, idk still weird