Machine Learning for 3D (p2 what is 3D?)

  Рет қаралды 3,325

IndividualKex

IndividualKex

2 ай бұрын

follow along: huggingface.co/learn/ml-for-3...

Пікірлер: 6
@TheLoneWolfling
@TheLoneWolfling 2 ай бұрын
Has there been any recent developments in storing gaussian splats in a somewhat space-efficient form, and/or in a manner that allows LoD of some form? Right now the state of the art seems to be "Here's this model of a bicycle and a few meters of the park around it that requires 24GiB of VRAM to render".
@ThomasPlaysTheGames
@ThomasPlaysTheGames 2 ай бұрын
Yes. There are at least 3 papers implementing some form of LoD. The most recent one is named "Octree-GS: Towards Consistent Real-time Rendering with LOD-Structured 3D Gaussians" Regarding storage optimization, Aras Pranckevičius has done some extremely impressive work.
@TheLoneWolfling
@TheLoneWolfling 2 ай бұрын
@@ThomasPlaysTheGames That's great; thanks for the pointers. Hopefully we see this work its way out of the land of tech demos sometime soon.
@ThomasPlaysTheGames
@ThomasPlaysTheGames 2 ай бұрын
@@TheLoneWolfling We'll get there in a few years. There are already demos for importing GS into unity & unreal; and an interesting guy has generated an importer for VRchat.
@manamana5495
@manamana5495 21 күн бұрын
Good video but why is it normal today that people no longer take the time to discuss things in detail? The video feels like you were stressed and had to hurry.
@can2835
@can2835 4 күн бұрын
imo better for my abysmal attention span
Machine Learning for 3D (p3 multi-view diffusion)
2:59
IndividualKex
Рет қаралды 3,3 М.
Why Does Diffusion Work Better than Auto-Regression?
20:18
Algorithmic Simplicity
Рет қаралды 242 М.
How Many Balloons Does It Take To Fly?
00:18
MrBeast
Рет қаралды 168 МЛН
Каха и суп
00:39
К-Media
Рет қаралды 6 МЛН
Looks realistic #tiktok
00:22
Анастасия Тарасова
Рет қаралды 105 МЛН
3D Gaussian Splatting! - Computerphile
17:40
Computerphile
Рет қаралды 124 М.
Reinforcement Learning from scratch
8:25
Graphics in 5 Minutes
Рет қаралды 49 М.
The moment we stopped understanding AI [AlexNet]
17:38
Welch Labs
Рет қаралды 761 М.
How to train simple AIs
12:59
Pezzza's Work
Рет қаралды 68 М.
What is Hugging Face? - Machine Learning Hub Explained
10:05
NeuralNine
Рет қаралды 10 М.
Why Computer Vision Is a Hard Problem for AI
8:39
Quanta Magazine
Рет қаралды 125 М.
A new way to generate worlds (stitched WFC)
10:51
Watt Designs
Рет қаралды 518 М.