For some reason, I genuinely thought image to 3-D was possible like eight years ago and set up several projects based on rapidly modeling all of the complicated parts using images. Going back and finding all of that material ready to go as the tools actually become available has been like watching a dream come true
I just saw the future. I remember being a kid, thinking, “imagine no TV.” Picture someone from the 1800s seeing a TV or trying to guess what would exist in the future. I would have never imagined a TV. So I wondered what we could possibly have in the future that I couldn’t even guess when I was ten. Sure, I own a drone and an iPhone, but I just got the feels and remembered being a kid, having these thoughts when I saw you spin the car in the forest. Sometimes it feels like nothing new is happening, but all this AI has been amazing in just one year. I can’t imagine what some of you will have in 50 years.
I genuinely feel that the main thing hold AI art back at the moment is less so the training of better models and more so the need for intuitive tool like this for more naturally interacting with the models.
Hey there! Diego from Krea here. Thanks for the shoutout! You’re right, no ad here hahaha… almost all money goes only for GPUs to support fast speeds of doing AI inference online! But for real thanks for the shoutout and many many many more updates coming in the next months!
It's very impressive with tons of potential. I hope in the future you can use your own 3D objects and have greater consistency between the object and the final version of it in the image. Still, it's an impressive start and I'll certainly be testing it out.
I'm 100% sure they don't use any proprietary models. You can do this locally with Nvidia Sana. The fact that the conversion to 3D took so long is a dead giveaway that the image generation speed comes from Sana and not from 20 GPUs working on your image.
You only touched the surface of what you can do with this tool. Great job, great video! Being able to enhance it then turn it into a video would have been nice.
It’s a great feature, and thanks for the quick tutorial and your genuine enthusiasm is really infectious. Did you test it with people/characters at all? Just curious. I will tomorrow as I’m actively using Krea as part of my pipeline on a short ai film and this looks like a great new add on.
Thank you Olivio, this is very interesting. I was just a bit disappointed when I took a second look as the car hasn't the same shape when you compare the views from the left and then from the right. I hope this will be fixed!
You can add in and create 3d people... but the face of the girl I used went from really pretty, to something that would have lived underground in Middle Earth!
@OlivioSarikas yea, I just searched some videos on it. It's a powerhouse. I'm gonna start using more now,, with 3d and this insane image editor, it could be the best out there. Midjourney new one is useless, forever getting flagged, censorship on steroids
Which model are they using to create the image gen? Doing localling is not possible? Its been doing before right with other programs, nothing new? Daz studio has this feature too, no?
This is very interesting development. By the way, I'm in Bangkok as well. I do a lot of voice work for animation, games along with film and TV work for Workpoint, Kantana, GMM, etc. It would be great to connect with your friend.
This is NOT an AD 🙂👍
and it ‘aint bad.
We don't mind advertising if the products are super cool!
can you convert an image of a human ?
I can't believe how fast the realtime generation runs. This is truly groundbreaking.
Yeah this is absolutely amazing. The potential here is wild. And then image to video comes on top of this also - astonishing what we can do with this.
I love how fast this makes scenes. This is the way to go for graphic novels. Imagine this with your own style lora
This is the future. You're painting with detail instead of lines. Art has evolved!
For some reason, I genuinely thought image to 3-D was possible like eight years ago and set up several projects based on rapidly modeling all of the complicated parts using images. Going back and finding all of that material ready to go as the tools actually become available has been like watching a dream come true
Thats freaking crazy! it is a major breakthrough for storyboarding rapidly for movies, animation, comic book panels, advertising, tv shows.
Very mind blowing, This is indeed the future. Thanks for demonstrating it
Pretty amazing. Wish there was something this fully featured that could be ran locally!
give Krita + Comfyui Krita Plugin a shot. it doesn't have 3D (i think) but it has painting + live LCM
Comfyui+ krita + trellis
I just saw the future. I remember being a kid, thinking, “imagine no TV.” Picture someone from the 1800s seeing a TV or trying to guess what would exist in the future. I would have never imagined a TV. So I wondered what we could possibly have in the future that I couldn’t even guess when I was ten. Sure, I own a drone and an iPhone, but I just got the feels and remembered being a kid, having these thoughts when I saw you spin the car in the forest. Sometimes it feels like nothing new is happening, but all this AI has been amazing in just one year. I can’t imagine what some of you will have in 50 years.
now imagine doing the same, but in VR :)
I genuinely feel that the main thing hold AI art back at the moment is less so the training of better models and more so the need for intuitive tool like this for more naturally interacting with the models.
Holy moly!! That's amazing. Thanks for sharing Olivio
Please do more videos of this.
Hey there! Diego from Krea here. Thanks for the shoutout! You’re right, no ad here hahaha… almost all money goes only for GPUs to support fast speeds of doing AI inference online!
But for real thanks for the shoutout and many many many more updates coming in the next months!
Subscribed!!
It's very impressive with tons of potential. I hope in the future you can use your own 3D objects and have greater consistency between the object and the final version of it in the image. Still, it's an impressive start and I'll certainly be testing it out.
the car changes a lot here, if you make a lua of the car it'll stay consistent
Absolutely insane in a good way, Wished there was a localized model.
This feature definitely raises the bar 🔥🔥
totally. this was the most fun i had with ai in a long time
Hi, Olivio!! This is what I was waiting for!
Yeah, this feels like experiencing the future
Awesome!
It's like an evolution of wire framing ... where the need for wireframing is completely made redundant.
Wow! Really mind blowing.
yes, i love this so much
krea is simply the best of the beast!
I'm 100% sure they don't use any proprietary models. You can do this locally with Nvidia Sana. The fact that the conversion to 3D took so long is a dead giveaway that the image generation speed comes from Sana and not from 20 GPUs working on your image.
This is awesome! I hope we’ll also see something like this released as open source.
check out krita with comfyui plugin
Impressive!
You only touched the surface of what you can do with this tool. Great job, great video! Being able to enhance it then turn it into a video would have been nice.
yes, totally. I just wanted to give a quick impression of what it can do, focusing on the 3d model generation and usage
The process speed is crazy! The fastest render time I have ever tried!
3d export?
That’s freaking awesome!
It’s a great feature, and thanks for the quick tutorial and your genuine enthusiasm is really infectious. Did you test it with people/characters at all? Just curious. I will tomorrow as I’m actively using Krea as part of my pipeline on a short ai film and this looks like a great new add on.
Looking forward to this moving offline and local with Trellis. The tools exist, someone just needs to package it together like this.
Krea has some cool features
So wild how fast that generates🤯
Wooooow!! This is sooooo goood! It’s amazing how fast it is. Can we import our own 3d models? And can it convert anything to 3d? Like a character?
Thank you Olivio, this is very interesting. I was just a bit disappointed when I took a second look as the car hasn't the same shape when you compare the views from the left and then from the right. I hope this will be fixed!
you can reduce the ai strength and then it looks more like the input, but this isn't like a flux lora of course
This is crazy mind blowing stuff. 😱🤩
so promising! can we try a character instead
character should be no problem too :)
How do you remove the background from your own portrait? Do you have any tools to recommend?
Bam! 💥.
(imagine setting up several sets of stills and then run a interpolation.)
that sounds like a cool idea :)
You can add in and create 3d people... but the face of the girl I used went from really pretty, to something that would have lived underground in Middle Earth!
yes, the faces are not detailed in 3D - but keep in mind that these models are just for the AI get an idea of the composition you want
havent they got a new editor, its like very different photo editor with clever prompting ? i'll have to check it out
They have a editor for images on a canvas, but this is for live generations
@OlivioSarikas yea, I just searched some videos on it. It's a powerhouse. I'm gonna start using more now,, with 3d and this insane image editor, it could be the best out there.
Midjourney new one is useless, forever getting flagged, censorship on steroids
We all know what you're going to do with this for the rest of the night...
Now, can we self host this
6:34 kann man sagen was der Kreis darstellen soll?
Which model are they using to create the image gen? Doing localling is not possible? Its been doing before right with other programs, nothing new? Daz studio has this feature too, no?
Looks like they are using trellis, which is fine. Can you export the model from krea?
Try this tool using a image of person, it is amazing!
what model do they use? is it some super fast 4 steps model? like schnell
WOW!! 🎉
that's insane
와! 이거 진짜 장난 아닌데요.😮
Superb! Krea is so innovative than Midjourney.
its cool but it keeps changing the car and not using the actual car you provided
I have just recharged Krea Max.🤣
awesome master
WOW
Can this tool turn a character (photo) to 3D?
wow so cool!
Krea is good.
absurdly syrreal stuff.
we are living in the future
insane
how about real time render in local? krita + Trelis?)
great video!!!
thank you. i had so much fun making this
Is it free or a paid subscription
paid, but totally worth it for this tech. if you want this free minus the 3D support, check out krita with comfyui plugin - i have a video on that
Is this free?
This is very interesting development. By the way, I'm in Bangkok as well. I do a lot of voice work for animation, games along with film and TV work for Workpoint, Kantana, GMM, etc. It would be great to connect with your friend.
if you want, let's meet. write me a email through my about page with the topic "meet in bangkok" - i am close to EM Sphere
but could you add people?
black magic
best video on this!
Real time kit bashing.
ФАНТАСТИКА
!!!
This is NOT a COMMENT
I used to love KREA until they pulled the plug on NSFW content. T'was a pity.