Topology refers to the way in which vertices, edges, and faces are organized to create the surface of a 3D model. Picture it as the digital DNA of your characters and objects.
I was pretty big into 3D modelling at one point in my life, I used to love to work in Maya and Zbrush creating characters. Nowadays I still dabble with a program called Nomad Sculpt, a cheaper more accessible version of Zbrush.I still have aspirations to one day in the future finish my game after I complete my masters degree...hopefully.
The idea behind this model is to help create detailed characters alongside their topological counterparts to assist 3d sculptures in visualizing how the simplest way to get complicated characters with limited polygons. But mostly I just thought it was a pretty neat idea.
I recommend keeping the weight between 0.8-1
If you choose to use other loras along side this one to get a particular character, I recommend keeping those loras below 0.8.
This works for both left to right and top to bottom depending on your chosen resolution.
Structure your prompts as follows:
"This is a digital image showcasing two different versions of a [character]. On the right(or top) of the image is [character description of your choice].
On the left (or bottom) side of the image, the [character] is depicted in a wireframe 3d model, showing its structure and intricate details. Topology."
IT DOES OBJECTS AND VEHICLES AS WELL AS WELL AS CHARACTERS.
You can also actually just leave the prompt as just "character" and Flux should spit out many different characters along with their topology, very consistently. If you just want to be inspired by a random design.
BUZZ TOWARDS THE WORK I DO IS GREATLY APPRECIATED. THAT'S ALL FOLKS!
Description
FAQ
Comments (11)
wow, very interesting! I used to do 3D modeling before I learned about AI)). But 3D is still interesting to me.Thanks!
Neat lora. Wonder how long it will be until we can port something like this into blender and have it generate a 3D model with AI.
Thanks.
The way this technology space is evolving, I would say fairly soon, there are already some text to 3d options out there. Only a matter of time.
@Wolf_Systems wwouldn't shock me if it were to happen within the next year or so. We already have concept art loras for games like csgo and valorant. Would be so cool to see some day the AI being able to generate fully unique levels or feed an already existing level to it for it to recreate it. But yeah I definitely see this happing within the next 3 years tops easily.
actually, it can be done rn, meshy is ok but this new model should assure you Deep Imagination Research | NVIDIA
@gemikawai09654 I've used meshy before, its pretty good, even textures the model, lets you download the obj. Havent used NVIDIA one
@gemikawai09654 for map design or character design? because I've always wanted it for mainly map design jusit so it would make my job way easier.
@ZooterTheWooter character design, it does pretty good meshes, and it can also texture them relatively well, you can then download the files and fix them up on your own. Not sure about level design.
you already can right now
check out trellis