DeepNude source released

github.com/deepinstruction/deepnude_official

Attached: logo.png (192x192, 24.62K)

Other urls found in this thread:

github.com/open-deepnude/deepnude_official
github.com/deep-man-yy/easydeepnude
github.com/deep-man-yy/pepenude)
bitbucket.org/deepmanyy/easydeepnude/src)
gitlab.com/deepmanyy/easydeepnude)
notabug.org/deepmanyy/pepenude)
gitgud.io/deepmanyy/easydeepnude)
huffpost.com/entry/deepfake-porn-heres-what-its-like-to-see-yourself_n_5d0d0faee4b0a3941861fced
phys.org/news/2019-06-deepnude-app-women-furor.html
vice.com/en_us/article/qv7zkw/create-fake-videos-of-faces-samsung-ai-labs-algorithm
vice.com/en_us/article/kzm59x/deepnude-app-creates-fake-nudes-of-any-woman
mic.com/p/the-deepnude-app-is-the-terrifying-future-of-deepfakes-18153256
twitter.com/NSFWRedditVideo

Did someone hack their servers or did they put the AI in the app, i.e. it worked offline? Is there something for end users?

You can tell it was made by a russian in babushka's basement

cant you just train it on a set of nubile jailbaits

Unless you have access to NSA's and FBI's stash, forget about it.
As it currently is, algorithms, or "deep learning" as the coffee house "programmers" call it requires a very extensive data set.

$ python main.pylibpng warning: iCCP: known incorrect sRGB profileExecuting phase: dress_to_correctExecuting phase: correct_to_maskTraceback (most recent call last): File "main.py", line 30, in main() File "main.py", line 21, in main watermark = process(dress) File "/home/ares/.tmp/dn/deepnude_official/run.py", line 115, in process model.initialize(opt) File "/home/ares/.tmp/dn/deepnude_official/gan.py", line 69, in initialize self.__load_network(self.netG) File "/home/ares/.tmp/dn/deepnude_official/gan.py", line 89, in __load_network network.load_state_dict(torch.load(save_path)) File "/usr/lib/python3.7/site-packages/torch/serialization.py", line 382, in load f = open(f, 'rb')FileNotFoundError: [Errno 2] No such file or directory: 'checkpoints/cm.lib'
help

go back to 4chan with this shit

fag

The source did not include the trained neural network libraries required to run the software, therefore a fork has been made to include those "pickles", it can be found here.

github.com/open-deepnude/deepnude_official

Useless until it can make anuses.

crashes with pretty much everything that isn't the example file, another quality python product

how's this shit popular? It barely works outside of some set, perfect cases. Horny 14 year old me with photoshop was a lot more efficient.

Attached: 1520468542.jpg (270x270, 26.79K)

It either fucks up the creation of the mask, or the size of the mask.
We should be able to fix that relatively easily, I think.

it had a few small issues that I needed to correct for my gentoo installation and the current libraries installed. The picture also has to be a 512x512 png. Also removed the "fake" markings.

Don't bother, it's garbage. Only works on bikini pictures when the woman stands in a very specific way. You really can do better with photoshop.

Post some results with originals so we can see without trying.

How much example file/"images" would it be?
I''m almost pretty sure that with all the porn floating around various website, there is enough videos to hae an almost complete dataset.

I don't know what the controversy was about with this. It's a very mediocre program. It fails spectacularly. Here are some "good" ones. I tried to get a mix of different races. Haven't tried body styles yet.

Attached: 1_on.jpg (512x512 82.91 KB, 288.44K)

Attached: 2_on.jpg (512x512 90.57 KB, 394.71K)

The lowest level of ML starts at 10k, but that's trash. Try a couple million data-points if you're aiming at using it with different races/clothes/positions/etc.

As said, this one's pretty trash tier, only really works on bikinis when the chick directly looks at you, and it gets weird quick if the chicks are too thin (or, as it's known outside of burgerland, not fat)

Attached: 3_on.jpg (512x512 132.41 KB, 251.95K)

...

Attached: 4_on.jpg (512x512 13.88 KB, 165.4K)

Attached: 5_on.jpg (512x512 176.19 KB, 412.01K)

Attached: 6_on.jpg (512x512 109.34 KB, 216.98K)

...

Attached: 7_on.jpg (512x512 67.13 KB, 159.24K)

Attached: 8_on.jpg (512x512 34.77 KB, 283.32K)

Attached: 9_on.jpg (512x512 83.22 KB, 330.65K)

Run this one through please.

Attached: 81Fq1AvKh0L.jpg (1769x2060, 462.24K)

Attached: 10_on.jpg (512x512 82.17 KB, 300.77K)

And sometimes it just fucks up

Hmmm... how do the CP laws work around this stuff. Dun want to get in trouble.

Attached: 12_on.jpg (512x512 86.73 KB, 311.35K)

You can google her she's 23.

Attached: Screenshot from 2019-07-06 07-25-08.png (452x382, 49.67K)

The archthot kinda works, sometimes.

Attached: likethe90s.png (512x512 423.85 KB, 514.1K)

Had to reduce the file size otherwise the program wouldn't handle it.

Attached: 13_off.png (512x512, 380.44K)

...

Well reduce the file size after changing it to png. I'm sure there was some better way of doing it. But your pic just crashed.

Just showing that the program sometimes works sometimes doesn't. Crashes without warning. Sometimes produces unbelievably bad results.

I appreciate your research, user. I actually didn't know it only worked on images with specific conditions met. Seems like whenever new software like this comes out people discuss it like it's perfect stable software, and not a very early build. Happened when Adobe showed off their audio photoshop, happened when that make-up remover app came out, also that real-time facial capture, etc. Reminds me of when anons first started to do x-rays, and the amount of idiots requesting "XRAY PLZ!!!" and posting pictures of women in multiple layers of clothing.

What happens if you feed it a picture of a woman that is already nude?

Attached: 157667.jpg (2883x4324, 662.6K)

Oh my!
Boner status: 11HARDNESS

Judge for yourself.

Attached: LovecraftianHorror.png (512x512, 302.03K)

Did you ever see any of TWDNE's creations?

Attached: example-67955.jpg (1024x1024 38.48 KB, 75.65K)

few more

Attached: example-68761.jpg (1024x1024 63.82 KB, 41.59K)

shit nigger we will be replaced

This is an SFW board, spoiler faggot

They should combine the tits AI with the dogs AI and have dogtits and titdogs everywhere.

Attached: figfan02b.jpg (661x1068, 64.06K)

Get out, underage faggot.

modern technology was a mistake

Not him, but you're an idiot. There are rules on this board you have to obey - the first is threads should be about Zig Forumsnology, the second is that this board is SFW, so porn, shit, or shitty porn must be spoilered. If you don't want to obey these rules fuck off to >>>/porn/

Interesting results

Attached: output.png (512x512 450.25 KB, 168.42K)

Wasn't there already a uncensore tool called deepcream for animu?

There is one last thing I forgot

Attached: mantest_off.png (1280x720 418.36 KB, 165.92K)

...

kek

Still looks better than the average tranny despite being the result of a hacky algorithm.

It makes me horny as a muscle girl fan.

this is hilarious

...

github.com/deep-man-yy/easydeepnude

Attached: ClipboardImage.png (1380x860, 1.15M)

This. I want to see assholes.

You underestimate my power

Github shut it down, do you have a mirror?

I'm not gonna be surprised if there'll be a Edward Snowden equivalent, except for Cheese Pizza instead of confidential three letter agency or military operations.
Weirder shit has happened after all.

Or just a neckbeard with decades of "archiving"

fag

Maybe someone can leak that Epstein collection.

Found the jew

Repo got purged

Can anyone upload the original files somewhere? Github and co are wiping it anywhere they see it.

dl thread
- deepnudes whatever released
>The original DeepNude Software and all its (((safety measures))) (((have been violated and exposed by hackers))). It no longer makes sense to hide the code. DeepNude uses an interesting method to solve a typical AI problem, so it could be useful for researchers and developers working in other fields such as fashion, cinema and visual effects.
What the fuck is this? What fucking """failsafes"""? How would that even matter when in 1990 there were already much better fakes (and even paintings before that would be more realistic). Let me guess, they had some code that makes a watermark, and someone commented it out or nopped it.
>If you followed our story you will know that we have decided not to continue selling DeepNude because we could no longer guarantee enough (((safety))). The original DeepNude app was intended to bu fun and safe: (((we knew our customers))), (((images were associated with them))) and watermarks covered the photos. (((But after 12 hours of launch))), due to viral articles and clickbaits, (((the software had been hacked and modified))). (((With multiple illecit DeepNude version in the web, anonymous and unknown users, virus and malware, the assumption of security dissolved soon))). (((There are no valid security systems, when hackers from all over the world attack you))).
>We could have made a lot of money, but (((we preferred to limit the misuse as much as possible))). If you have enjoyed our work or you would like our research not to stop there, make us a donation.
Okay, so all they had was a watermark, as I predicted. It sounds like the code was not public before and you had to use their shit through a server? So what? It wasn't groundbreaking in any way.
This entire thing reads like some idiots trying to feed the narrative that "oh no computer has too much power for you goy think of the women. you need to have every action you do online recorded to make sure you don't do misuse". Not to mention it feeds the retarded narrative about """AI""".
And

Who could have foreseen Microsoft removing this project? WHO?

As far as I know Github did not close the repository, it was closed by the owners of DeepNude, I said this because they sent me an email warning me of a DMCA and in that email they indicated that they had closed their repository.

Good thing I downloaded it, it was a pain to get it run though and the results are disappointing, especially with teen girls the real ones are much better.

You wouldn't need any of this la deep algorithm shit if you knew how to paint nigger.

galaxy brained tbh

They say every artist has a folder that's been painted one-handed.

Somebody be brave and change the course of history and post a link.

Seriously though, wtf where are the mirrors?

Attached: 1561933022859.jpg (400x388, 39.88K)

kys

Attached: 43.jpg (480x360, 12.2K)

I read that as minors.

The wish was father of the thought.

It looks like that repo is dead. I downloaded zip file a few days ago, and it lists these mirrors in README:
- [Github](github.com/deep-man-yy/pepenude)
- [Bitbucket](bitbucket.org/deepmanyy/easydeepnude/src)
- [Gitlab](gitlab.com/deepmanyy/easydeepnude)
- [NotABug](notabug.org/deepmanyy/pepenude)
- [GitGud](gitgud.io/deepmanyy/easydeepnude)
Only the last two are working right now.

Attached: 1549231911570.gif (400x560, 597K)

DeepCreamPy is not made for clothes, it is made for uncensoring black bars

the funniest part is they thought they needed a watermark for us to know the generated pictures aren't real

This is the next western retardation after jailing people for "child porn" because they have a picture of a 17 year old girl in some folder on their computer.

GOT A LICENSE FOR THAT COMPUTER, GOY?

huffpost.com/entry/deepfake-porn-heres-what-its-like-to-see-yourself_n_5d0d0faee4b0a3941861fced

phys.org/news/2019-06-deepnude-app-women-furor.html

vice.com/en_us/article/qv7zkw/create-fake-videos-of-faces-samsung-ai-labs-algorithm

vice.com/en_us/article/kzm59x/deepnude-app-creates-fake-nudes-of-any-woman

mic.com/p/the-deepnude-app-is-the-terrifying-future-of-deepfakes-18153256

is it possible to modify it so it works for children?

I want to see all feminists executed.

stop waiting for things to happen, make them happen

are there any actual real looking fake pronz? i just wasted 2 hours looking through "deepfakes" on a few websites, they're all 100000% bullshit. even the decent looking ones look nothing like the person they are imitating, and more like a mix of the original and new face. but what would you expect when these shitty algorithms are just an extrapolation of a bunch of data with no regards to any sort of physics or other aspect of reality

Mfw NVIDIA hasn't updated the driver since March 2018.

Attached: fug.jpg (480x360, 9.84K)

This.

no, its a meme. The outrage is fabricated, like usual. You can do more better with photoshop

fuck deepfake porn
install Tor and download real porn, child porn. it's best porn

maybe some day this will be so good that you cant know if your cp is fake or real.

Can someone please run this image through?

Attached: 21d04ca09fc404e725f90220344a8d1430f15c286ae9b99c554a26fbfd51a5bd.jpg (759x823, 139.78K)

kys dumb phoneposter

loli nudes next!

Can't you just invite her for a lunch?

she's dead

Ok this seems to be a problem. Then just download and install the software, nothing complicated.

That means it's even cheaper, all you need is a shovel and the gas to get there.