NVIDIA GET3D: AI Model to Populate Virtual Worlds with 3D Objects and Characters

Look up "AI for optimizing game meshes". We need AI to make to bake down Insanely high poly meshes on to a AI mesh with AI made LODs.
You'll find the Nvidia forums.

I just want a digital Earth twin as the map of a game like GTA.

¿alguna forma online?

it's completely unethical and fascist to treat artists of their many years of work as fodder for the AI ​​monster. Nvidia should be boycotted

Is NVIDIA GET3D Can make humanoid 3d model such as "Korea girl" produce 3d model by AI?.

can this be implemented inside a video game so that the inside the game ai generates new things and make the game feel fresh everytime with new things added to the scenes...

Fck Nvidia they abandoned Nvidia 3D Vision driver support leaving me with a lot of PC gear redundant now so why would i go for their next poorly advertised 3D product they should have promoted it in stores where people could try it out it infact it was incredible as long as you can tolerate stereoscopic 3D,i loved my 3D Vision all we want is driver support for our old gear that cost lots of money and still works if given the support.

Why not use AI for more important things like solving diseases. saving lives and not taking away people's hard work and their lively hood.

This is BS i worked my ass off for years to get good at 3D, and now I'm going to probably lose my job in the near future. Please please please stop. I love what I do.

Yea I hope atleast 3D art stays safe for like 2 years. AI is the worst thing to have happen to the internet.

You know what I would find really cool? If at some point AI could create a whole animated movie for you. :D
Because I can't animate...
Although I wouldn't want AI to do it ALL for us... the script writing... the animation... at some point we might have a self-creating entertainment industry. And what then? Will that feel hollow? Will we feel that the human spirit is missing in those works? Will we notice? Interesting times. Hopefully fun times too. Me, awaiting further developments with a certain trepidation, but also a lot of curiosity : )

Looks pretty basic, but I hope this gets developed further and there can be a Midjourney for 3D art.

Full 3D Generation in 2023 soon

I love making Ai art & this is next level! I can't wait to get my hands on Get3D!

Enter the Matrix soon

Just mixer w filter

mfw i spent 8 years to become a 3D environment artist and now my job security is in jeopardy... It's not long before they'll be able to generate game topology/quad meshes..

Welp, we lost Concept artists, Texture artists, 3D modellers in a span of 3 months, who's about to get fucked next?

I knew this was going to happen

IS THERE AN INSTALLATION TUTORIAL ANYWHERE PLEEEAAASEEE ?!?!?! (and thnx byeee!)

Please, stop taking our jobs away. Thanks.

It is clever and will doubtless get much used. Especially by the younger generation who seem to like and want easy one click results. For me, the making is the point of life, be it in real world crafts or digital. But I obviously have no place in this changing world at my advanced age, left school before digital pocket calculators existed. I wish the young well but it looks like a poor way to be from my perspective. What will you do with your lives? 'Flower arranging' ai made stuff? Until ai does that for you too?

I fucking need this!!

ХУЙНЯ! Цена на Четырёх Тысячную линейку ЗАВЫШЕНЫ ОХУЕННО!

The more we venture into the metaverse, the more it will auto-populate. It has infinite possibilities, infinite opportunities.

I want it !!

Nice!

Oh, 3d artists are soon gonna be out of their jobs lol. Just like 2d artists too. AI gonna replace them all(well, most of them, for generic tasks, anyway. Most low tier companies will be fine with generic objects in their games for saving the costs). And AI will only be getting better.

This will change everything

😎😎😎😎😎😎😎😎😎

Thanks 😢

This is the start of taking real world partial data about a location like cellphone photos and having it crank out a high fidelity simulation of that place in time. We should be able to explore vast swathes of the real world in vr in a few years

GTA 8 will be amazing

The way I read it i got that the model is not trained but it can be trained with any subject of datasets. My question is how is this out-of-the-box? What can I do with it if i manage to run it? I hope someone comes with some tutorials for those of us who are not developers or researchers! This is from NVIDIA blog article :

" GET3D can generate a virtually unlimited number of 3D shapes based on the data it’s trained on. Like an artist who turns a lump of clay into a detailed sculpture, the model transforms numbers into complex 3D shapes.

With a training dataset of 2D car images, for example, it creates a collection of sedans, trucks, race cars and vans. When trained on animal images, it comes up with creatures such as foxes, rhinos, horses and bears. Given chairs, the model generates assorted swivel chairs, dining chairs and cozy recliners"

Can it make me?

Humanity needs to chill before we make our selves obsolete. Like I get this won't 100% replace artists. We said the same thing about trains vs planes and we said the same thing when it came to robots entering factories. In 10 years time we will be asking where version 3.0 is and if it can generate a level faster than 2.5

Bravo to NVIDIA researchers!!! You guys are amazing!!! Can’t wait to use this product

Okay. I guess I'll just give up animation and modeling. What's the point if a machine can do it better and faster. If you welcome this then bend over and say aah to the new overlords. Nice knowing you all.

Cooool, so how long until we can get a 'Panda Cop riding a Rhinocycle'?

Nvidia needs to operate in Mexico, very cheap labor there

@Nvidia Developer , When would we be able to generate a 3d model by describing it, much like a 3d version of Dall-E2?

where can i find this software ?

teey took errr j0bssssss

they terk err jerbs!

source: 3D artist.

What gpu i need and how can i make it on my own PC?

Based on that video, I'm not altogether sure it delivers on it's sales pitch. That's some really nasty geometry and kooky looking results.
But it's a start in at what is quite a tricky problem. There just isn't the wealth of semantic object data that there is for images.

all for the price of your left kidney

WOW