Create Anime Characters with A.I. | Hacker News Reader

Create Anime Characters with A.I.

make.girls.moe

274 pointswei_jok48 comments

It's copy-pasting parts of the training images over and over.

In figure 8 of the technical report [0], compare the hair in images (0,0), (2,0), (3,0), (3,3), (4,4).

The paper suggests their method generates copyright-free images, yet they are very obviously derived from the input images and you can identify the parts of individual input images that are mashed together to form the output.

All in all their method seems to be performing "obfuscated memorization," in the sense that the generated images are scrambled up enough to fool their plagiarsim-detector loss function.

[0] http://make.girls.moe/technical_report.pdf

You can also view Figure 8 online: http://make.girls.moe/news-img/fixed_noise.jpg

But as the online article states, that figure represents a case where the model is explicitly set to "generate images [which] have similar major visual features with different attribute combinations": http://make.girls.moe/#/news

So some degree of repetition is to be expected, since you've turned off random noise. And despite that the images do still exhibit some variation if you look closely.

That is an awesome choice of domain name ;)

For the uninitiated: (form Wikipedia) Moe (萌え, pronounced [mo.e]) is a Japanese slang loanword that refers to feelings of strong affection mainly towards characters in anime, manga, and video games.

http://nic.moe/en/

This is what the domain name was created for IIRC.

AI anime character generators, of course.

I have a .moe domain, partly because of my hobbies and also beacuse it was surprisingly cheap. Unfortunately szyslak.moe was taken :)

I just seem to end up generating pictures that look like smashed crabs: http://imgur.com/4PbYuEt

Or a perfect villain

Maybe it's the hidden heterochromia feature

And short hair seems to produce male characters

> And short hair seems to produce male characters

Unsurprisingly! They didn't exclude male characters, so if you ask for short hair, you'll be more likely to be drawing from a male-biased area of the latent space/noise.

Is there a reason it doesn't generate images as pretty as the samples on the news page? (http://make.girls.moe/news-img/samples.jpg)

By the way, the Getchu and illustration2vec links on the news page are broken.

Edit: This part from the Tips page might be why it initially didn't generate great images:

The input of the model consists of two parts, the random noise part and the condition part. If you generate a good image, you could try to fix the noise part and use random conditions to get more good images. We have observed that a good random noise is important for the better generation.

Edit 2: Actually, no. According to the news page, if the noise is fixed, the generated pictures would be all similar.

Wow, the quality of generated face is pretty nice. It's not ugly as I usually see of GAN.

http://imgur.com/a/A5gID i sometimes get this moetron monstrocity :P

Amazing this works entirely inside the web browser.

Yeah, came to ask--so the GAN model is loaded into the Javascript engine and run locally?

I thought DNN models are fucking huge?

From the "news" page:

In order to make our model more accessible, we build this website interface with React.js for open access. We also make the generation completed done on the browser side, by imposing WebDNN and converting the trained Chainer model to the WebAssembly based Javascript model. For a better user experience, we would like to keep the size of generator model small since users need to download the model before generating, so we replace the DCGAN generator by SRResNet generator can make the model 4 times smaller. Speed-wise, even all computations are done on the client side, on average it takes only about 6 seconds to generate a single image.

It gave me a warning to connect to WiFi and not load the site over a mobile data connection. So yes.

Oh, I didnt get that, would have been nice.

Not every.

In Cromium 46 (released in 2015) there is an error:

> Error during service worker registration: DOMException: Only secure origins are allowed (see: https://goo.gl/Y0ZkNV).(anonymous function) @ main.4188687f.js:1 descriptor_runner_webassembly.ts:172 ErrorEvent {isTrusted: true}worker.onerror @ descriptor_runner_webassembly.ts:172

> Uncaught RangeError: Invalid array buffer length math.ts:55 Model loading failed for webassembly backend. Trying next backend: Uncaught RangeError: Invalid array buffer length

In Firefox 45 (released in 2015) there is an error

> InternalError: uncaught exception: out of memory

(the necessary amount of memory is not reported)

Well, Chrome updates automatically and it's obviously not safe to run an outdated browser. So you have a bigger issue than a random script not working on your browser.

Looking at current browser market shares on netmarketshare.com (first link that came up via google search) it looks like this version of chrome accounts for 0.07% of traffic this month. Why are you pinned on such an old version?

Interestingly enough, the older version 45 has 6.08% share. Is some operating system pinning chrome 45 and perhaps providing security updates?

The paper, http://make.girls.moe/technical_report.pdf

Props for the domain name.

The algorithms generates very obvious and non-aesthetics artefacts: http://imgur.com/a/hxBop

Shouldn't it generate different avatars everytime? I get the same picture when I generate a new one with same non-random params.

If you've selected the 'Noise: Random' option, it should, yes.

Create anime girls. Doesn't anime have boys?

Well, "make.girls.moe" is the URL and name of the tool.

According to the technical report [1], they used character portraits from Getchu [2] for training data. A cursory glance shows that the overwhelming majority of the characters are female. As a result, the characters the tool generates are likely to appear female to our eyes.

After the code is open sourced, perhaps someone should try to create make.boys.moe using character portraits from otome games [3].

[1]: 12.8 MB PDF, http://make.girls.moe/technical_report.pdf

[2]: example from report (possibly NSFW?), http://www.getchu.com/soft.phtml?id=933144

[3]: games with a female protagonist and many male characters, https://vndb.org/g542

are you sure none of them are boys? :^)

Maybe some of the girl programmers that Google hasn't hired should make a site that generates sexy anime boys?

Probably a matter of dataset.

You don't need much of an A.I. to create the ordinary high school student :)

It generated a boy for me once.

Please don't propagate that sexist, binary view on gender. If you make the effort to criticise non-equal treatment, please always also include genderless, non-binary bigender or trigender, pangender, trans woman, trans man and any other-gendered.

I'd like to assume it's a sarcasm, but not aure I can.

i genuinely cant tell if satire. A+

So sad it has become this way.

;)

Yeah but 95% of HN are white heterosexual males so there you go

The remaining 10% are non-white but still heterosexual males.

10%...?

Shit, we're over capacity, this is against fire safety regulations!

I wanted to create myself (white male). :/

How about trap generator?

Draw a girl, call it a boy.

True. Just change URL to make.traps.moe

Profit.

Reply on news.ycombinator.com