I created an account and there seems to be a daily email to get 5 more credits. I can’t say any of my creations over the last few days have been particularly noteworthy, though. I had better luck with Dreamscope in the past.
I haven’t yet. There’s a bunch of ways to get free credits, and so far I’ve made well over a hundred images without having to pay.
Thanks, how do go about getting free credits?
I think I just got 5 free credits after signing up. (Not so much signing up, as connecting with Facebook).
Then, each day I can click on the alert in the upper right corner to claim 5 more.
FWIW I did find some downloadable pre-trained models online. E.g. there is a Wikiart one with 81444 images. If you have a GPU you can just run it yourself all day long…
Or you can go to
and just keep refreshing until you’re tired of it.
Like others said, you automatically get 5 per day (though you have to log in and “accept” them each day). Also, you get free credits for various accomplishments like getting a certain number of “likes” on your own creations, liking others’ creations, publishing certain numbers of your creations, and posting them to social media. If you have any social media accounts (at least the big ones – Twitter, FB, or IG) I think this is the most reliable way to get free credits. Up to 10 times per day, you get 3 credits per social media post if you post one of your creations and tag the NightCafe account in it. So that’s 30 credits per day by posting on social media. I happened to have an IG account that I never used for anything else that’s now full of my NightCafe creations.
So, I got curious about what sort of library of training images they used, so I gave it a few descriptions of famous images.
“Lena” didn’t turn out too well, and doesn’t at all resemble the image used for years in image-processing examples. It does appear to include a bare breast that it picked up from somewhere… Maybe it knows that the “Lena” image was cropped from a Playboy centerfold?
“The Pillars of Creation”
Vaguely similar to the Hubble picture of the Eagle Nebula (the “Cosmic” style might also have contributed to that), but the broad shape is completely different, as are the colors.
“13 men at a seder”
The Last Supper, it ain’t. The men appear to be in Orthodox Jewish garb, which I suppose makes sense, but there are a lot less than 13 of them.
“Portrait of a smiling noblewoman”
…I kind of like it. Again, definitely not the Mona Lisa, and the lady in question is wearing way too much jewelry and makeup, but that might be realistic.
“God and Adam touching fingers”
Those are some weird-ass fingers (or maybe weird ass-fingers), but the influence of Michelangelo here is unmistakable. Count this one as a hit.
“Two bored little angels”
Maybe a little Rafael here? But these angels look more pissed than bored.
“Self-portrait in a reflective orb”
Looks at least reminiscent of Escher, though I have no idea what that is in the reflection.
And given the name of the site, “A corner diner late at night”
Looks reasonable, but not at all Nighthawks.
I tried linking to my unused instagram account, but it says I have to have 50 followers which I definitely don’t. I’m not even sure why I have 3 followers.
I bought $20 a while back but, with as many credits as you get from sharing and rating and the five a day, I’m still around 150 after a lot of usage. No regrets though, I’ve spent $20 on video games I got less time and entertainment out of and it’s nice not to feel crunched if i want to enhance or enlarge something.
Put in Golden Gate Bridge | fog and got a scene out of Star Trek IV.
^^
Looks a bit like a William Turner piece.
Oh, and I think what happened with my “God and Adam touching fingers” image was that the initial library contained both the entire picture, showing God and Adam’s full bodies, and the close-up on just the fingers. The algorithm knew both were relevant, and tried to reconcile them by showing the fingers, but with the textures of the muscular bodies.
I thought I would compare “Brain in a jar” (horror) (artistic)
to
“Brian in a jar” (horror) (artistic)
As I evolved the Golden Gate Bridge scene, it turned into more of a waterfront. So I’m running with that in a Cyberpunk setting, will post pic after a few more evolutions.
Cyberpunk Waterfront
Wow, these are pretty good. Too bad they’re kinda small for desktop background use.
Brian looks remarkably happy to be in there, all things considered.
Hovercar Racing
Now THIS is podracing! [/Anakin]