@bingyikang @huggingface @_akhaliq @pcuenq @NielsRogge Wow. I’m curious how the demo videos were created? as in terms of visualising the reconstruction as it roams around the building, or the first video clip—amazing!
English
cailean
165 posts
cailean
@caileanfinn
Creative Technologist & Media Artist
Ireland Katılım Ekim 2021
384 Takip Edilen325 Takipçiler
Paper: arxiv.org/abs/2511.10647
Project page: depth-anything-3.github.io
Code: github.com/ByteDance-Seed…
Hugging face demo: huggingface.co/spaces/depth-a…
Thank @huggingface @_akhaliq @pcuenq @NielsRogge again for your support.
English
After a year of team work, we're thrilled to introduce Depth Anything 3 (DA3)! 🚀
Aiming for human-like spatial perception, DA3 extends monocular depth estimation to any-view scenarios, including single images, multi-view images, and video.
In pursuit of minimal modeling, DA3 reveals two key insights:
💎 A plain transformer (e.g., vanilla DINO) is enough. No specialized architecture.
✨ A single depth-ray representation is enough. No complex 3D tasks.
Three series of models have been released: the main DA3 series, a monocular metric estimation series, and a monocular depth estimation series.
The core team members, aside from me: @HaotongLin, Sili Chen, Jun Hao Liew, @donydchen.
👇(1/n)
#DepthAnything3
English
Now wait for a few more years(hopefully not that long) when robotics come to the same conclusion 😂
Stanford NLP Group@stanfordnlp
The final admission that the 2023 strategy of OpenAI, Anthropic, etc. (“simply scaling up model size, data, compute, and dollars spent will get us to AGI/ASI”) is no longer working!
English
imitations; simulations; emulations
AK@_akhaliq
Google presents Diffusion Models Are Real-Time Game Engines discuss: huggingface.co/papers/2408.14… We present GameNGen, the first game engine powered entirely by a neural model that enables real-time interaction with a complex environment over long trajectories at high quality. GameNGen can interactively simulate the classic game DOOM at over 20 frames per second on a single TPU. Next frame prediction achieves a PSNR of 29.4, comparable to lossy JPEG compression. Human raters are only slightly better than random chance at distinguishing short clips of the game from clips of the simulation. GameNGen is trained in two phases: (1) an RL-agent learns to play the game and the training sessions are recorded, and (2) a diffusion model is trained to produce the next frame, conditioned on the sequence of past frames and actions. Conditioning augmentations enable stable auto-regressive generation over long trajectories.
Français
If you guys want to install WHAM on windows, I added the instructions on how to do it here
github.com/yohanshin/WHAM…
and the blender importer here
github.com/yohanshin/WHAM…
I want to make a video tomorrow on how to do it in google colab, we'll se if I'm be able to do that 😀
English
@lmk_nathan42915 Hey! I was wondering if you have the link for download? 🙏
English
In what ways are you using #AI on your RTX GPU in your personal or professional life? 🤔
Reply with your answer and use #AIonRTX for a chance to win a @MSIGaming GeForce RTX 4080 GAMING X TRIO GPU. 👇
English
@carlosedubarret @toyxyz3 Is it possible to get accurate Z-Axis data? or is that just a limitation of converting 4D -> Blender?
English
@carlosedubarret Did you resolve the issue with the Z-Axis? Looks smooth af 😎
English
CEB 4d Humans, coming soon for Patrons
This is my atempt to make simple and easy to use the amasing code from
github.com/shubham-goel/4…
English
CEB 4d Human is arriving....
Doing some tests to release the early version for Patrons
English
@carlosedubarret @poomnattawat1 Ah yes! I was looking in that folder last night. Their track.py file is a direct copy of the main fn found in 4dhuman. I’ll try this and see
English
@caileanfinn @poomnattawat1 Use a debugger to run, it will make much easiert to find and fix the problems
English
Comparing SLAHMR (on the left) and 4D Human (on the right)
Processing 240 frames
-------------------------
SLAHMR: 4 hours
4D Humans: 15 minutes
For this video I tweaked 4d human result, removing the translation on Z axis
English
@carlosedubarret @poomnattawat1 How do you install SLAHMR with 4dHuman? I’m able to install both separately. As I see in their latest update they support 4dh, but I’m not sure how to put both together. Or am I misunderstanding how SLAHMR works? Thanks for the sick work!
English
@poomnattawat1 I just have my notes that I took while installing it.
Basically it was:
1-install detectron first using torch==1.13.0 and cuda 11.7
2-follow slahmr instructions
3-get 4dhuman installation and put near doirdslam
4-tweak a lot of things. A LOT OF THINGS ☺️, at least on windows.
English
@dvsch I don't think he is ignoring/disregarding current issues that plague AI with this post. He is just thinking on a grander timescale - nothing wrong with that imho. I'm sure he has talked about issues before such as 'bias' in his countless talks, and has a good understanding of it
English
@caileanfinn he’s just piggybacking on all the other dumb dumbs who are obsessed with AGI but ignore actual problems in bias today
English
Lex might talk to intelligent people but he’s one of the most gullible idiots I’ve ever seen. That people are only now seeing this is embarrassing. Man’s been joe rogan in a suit for years.
Christian Findlay@CFDevelop
WTF Lex Fridman... This guy talks with a lot of intelligent people, but this demonstrates a lack of understanding. LLMs do not offer any evidence that we are closer to creating sentience at all. Actually, they just prove how little we understand about sentience
English
cailean retweetledi
There are a bunch of papers following this one. NeRFs soon will be moving and all sorts of interactive stuff.
Like moving a car around a street kind of weirdness.
Alex Bergman@alexwbergman
We are excited to present our work at #NeurIPS22 on Generative Neural Articulated Radiance Fields! In this paper, we build on top of the capabilities of existing 3D GAN architectures and propose the first method for generating full human bodies! (1/7)
English
cailean retweetledi
Finally ML has managed to match the SHRDLU system that Winograd wrote in 1968. It took over 50 years but now it works in the real world and is learned from data.
Pete Florence@peteflorence
"Interactive Language: Talking to Robots in Real Time" interactive-language.github.io - Real-time, interactive, open-vocabulary, language+pixels -> actions - A new scale (~600,000 traj.) for language-conditioned behavior - Dataset, sim, models, code all to be released! (1/n)...
English
cailean retweetledi
Creative Machine v1.0 🤖 is now out and available in PDE!
As my Google Summer of Code ‘22 project with @ProcessingOrg, I developed a new Machine Learning library for beginners in ML and coding.
#gsoc #ml #MachineLearning #processing #creativecoding
English
What #Stacks projects are you excited about or you have dropping soon? Lets hear it!
English