Joed

Working with research is nice, but it is much better to research something that will be used to save lives. The following video is part of my work done in @SAFE_UP_ at @th_ingolstadt. #ros2 #imgui #computervision #PyTorch #cpp

@wenzeljakob Thanks for making such awesome libraries. Mitsuba3 and nanobind are my favorite open source projects.

Dr.Jit+Mitsuba just added support for fused neural networks, hash grids, and function freezing to eliminate tracing overheads. This speeds up optimization & realtime jobs and enables custom Instant NGP and neural material/radiosity/path guiding projects. What will you do with it?

Happy NN Christmas to all 😊 Here is an Enconder Christmas Tree 🌲🌲 #python #ai #neuralnetworks #xmas #MerryChristmas done with github.com/joedlopes/nn-s…

Very pedagogical. Must read! Free to read online version: bishopbook.com

I've just made neural network simulator from scratch in python :) youtu.be/2GsTY-X6W-Q?si…

@pmddomingos @elonmusk Does "better" mean LiDAR can replace vision or should it be added to the stack? You may ask @elonmusk what are the results of vision vs. LiDAR at heavy fog or rain for detecting objects (especially surfaces w/ specular reflection props.) at 15+ meters range.

@_jo_ed_ @elonmusk Works better than vision.

What @elonmusk is missing when he says that cars only need vision is that vision doesn't work without touch, and lidar is cars' equivalent of it.

@skalskip92 github.com/pjreddie/darkn… 2017 The code is so neat that you need 0 comments.

@_jo_ed_ Darknet made me cry so many times :/

YOLOv6 (2022) Bo Zhang et al. - paper: arxiv.org/abs/2209.02976 - code: github.com/meituan/YOLOv6

@skalskip92 Great review! I believe that part of the success of yolo is because it was open source, very easy to use and train. darknet's C + CUDA implementation is a beautiful piece of art.

oh, look! my YOLOv9 finished training!

I'm playing with DL and human body models, and #krita is quite nice for brainstorming

@karpathy @karpathy, in a non-distant future, you can spawn an AI NPC that is the clone of a human, who would you invoke to help and discuss with you?

Not me jealously looking at all the people getting their Apple Vision Pro today... I woke up to order mine a few days ago at 5am too, but I selected mail delivery instead of pickup, and it only tells you after you order and pay that this moves your time from Feb 2 -> Feb 6. And you can't change the delivery type later, even if you call customer support. I've been excited about AR/VR for a long time and I've bought every. single. headset. that has come out over the years. I haven't been converted to a regular user of any of them just yet, but I have no intention of stopping because one day it will be amazing. Forget image generation, we'll be generating entire synthetic worlds, and hang out in them with friends and AI NPCs. I wrote a silly post from back in 2017 expanding a bit more on this obsession karpathy.medium.com/virtual-realit… ok, i wait🧘

@HildeKuehne Now I am waiting for HOSVD time to come

I never thought there would come a day when people would dare to consider SVD to be anything but Singular Value Decomposition …

@CSProfKGD That's inspiring

It's ConvNet week in my #DeepLearning in #ComputerVision course.

A classic by @pointinpolygon @AndrewGlassner et al. The reading is super nice, I want to have time also to read all the reference papers of each chapter 😀

@augmentaltech That's a great project. I imagine in the future you can customize gestures and improve the user experience.

Today, we unveil the first public demo of the MouthPad^: smart mouthware for hands-free computer control. Witness Corbin swiftly navigate Google Maps on his phone using only subtle tongue movements. Cursor control, clicks, and swipes: full access.

Quaternions are nice😀 #python #programming #opengl #qt #pyside6 #pyqt

Playing with Euler Angles and Gimbal Lock before implementing a quaternion lib #python #opengl #computergraphics #cg #programming

@OdinLovis Great work. Is this a UV map (like dense pose) or just segmented parts?

if people want to test my Dense flow, here you go ;)

First time using Magic Animate at @dogstudio ! I created my own Dense Pose using Cinema 4D, I used ComfyUI , SDXL and controlnet, to generate visual. And finally MagicAnimate for the result. I think face are not good yet, but body are pretty interesting ;) !

@MattNiessner Great work. Now we can have infinite data to train and validate an eye tracking system.

(1/2) Check out 𝐆𝐚𝐮𝐬𝐬𝐢𝐚𝐧𝐀𝐯𝐚𝐭𝐚𝐫𝐬: Photorealistic Head Avatars with Rigged 3D Gaussians! We create photorealistic head avatars by animating 3D Gaussians on a parametric face model - edited and rendered in *real-time*! youtu.be/lVEY78RwU_I shenhanqian.github.io/gaussian-avata…

@ericnantel2 If you like vehicle dynamics you may check link.springer.com/book/10.1007/9…

@_jo_ed_ Also I think the rotation depends on the differential but I am not really a connoisseur of car physics.

Getting ready for implementing a vehicle dynamics lib. Rotating tires independently is also a nice way to teach the composition of multiple homogenous transformations :-) #python #opengl #computergraphics #vehicledynamics

@ericnantel2 Yes, you can implement a differential too. Here, I am just rotating the wheels, which are not well centered xD. We can implement in Mrald in the future :D.