sudhirPyadav's profile picture. I speak what I feel is true -- Open to change mind and discuss on anything. My views keep changing with information and maturity of my mind.

Sudhir Pratap Yadav

@sudhirPyadav

I speak what I feel is true -- Open to change mind and discuss on anything. My views keep changing with information and maturity of my mind.

Sudhir Pratap Yadav reposted

100M Gaussians. Streamed on a single RTX 3090. Welcome to @NVIDIAAIDev 's fVDB Reality Capture — bringing 3D Gaussian Splatting, volumetric data, and OpenUSD into one pipeline. 🎥 Full demo → youtu.be/ZnhBGmHyJqM #NVIDIA #GaussianSplatting #3D


Sudhir Pratap Yadav reposted

😠💢😵‍💫Tired of endless data collection & fine-tuning every time you try out VLA? Meet RDT2, the first foundation model that zero-shot deploys on any robot arms with unseen scenes, objects & instructions. No collection. No tuning. Just plug and play🚀 Witness a clear sign of…


Sudhir Pratap Yadav reposted

What if robots could improve themselves by learning from their own failures in the real-world? Introducing 𝗣𝗟𝗗 (𝗣𝗿𝗼𝗯𝗲, 𝗟𝗲𝗮𝗿𝗻, 𝗗𝗶𝘀𝘁𝗶𝗹𝗹) — a recipe that enables Vision-Language-Action (VLA) models to self-improve for high-precision manipulation tasks. PLD…


Sudhir Pratap Yadav reposted

Hot take: DAgger (Ross 2011) should be the first paper you read to get into RL, instead of Sutton's book. Maybe also read scheduled sampling (Bengio 2015). And before RL, study supervised learning thoroughly.

shaneguML's tweet image. Hot take: DAgger (Ross 2011) should be the first paper you read to get into RL, instead of Sutton's book. Maybe also read scheduled sampling (Bengio 2015). And before RL, study supervised learning thoroughly.
shaneguML's tweet image. Hot take: DAgger (Ross 2011) should be the first paper you read to get into RL, instead of Sutton's book. Maybe also read scheduled sampling (Bengio 2015). And before RL, study supervised learning thoroughly.

Sudhir Pratap Yadav reposted

Basic Sim2Real pipeline for Single leg🫡, trained using @kevin_zakka's MJ Playground. Basic reward just to track height in sim, no sensors in real life (ik, too basic). Would develop this further but for now shifting to DeepRobotics Lite3. If successful, would send PR to MJP.


Sudhir Pratap Yadav reposted

friendly reminder that I have a repo full of tutorials for vision language models, computer vision and even Omni models from fine-tuning to distillation

mervenoyann's tweet image. friendly reminder that I have a repo full of tutorials for vision language models, computer vision and even Omni models
from fine-tuning to distillation

Sudhir Pratap Yadav reposted

Simulation drives robotics progress, but how do we close the reality gap? Introducing GaussGym: an open-source framework for learning locomotion from pixels with ultra-fast parallelized photorealistic rendering across >4,000 iPhone, GrandTour, ARKit, and Veo scenes! Thread 🧵


Sudhir Pratap Yadav reposted

🚀 Introducing SoftMimic: Compliant humanoids for an interactive future — bringing humanoids into the real world 🤝 🔗gmargo11.github.io/softmimic/ Current humanoids collapse when they touch the world — they can’t handle contact or deviation from their reference motion 😭 😎SoftMimic…


Sudhir Pratap Yadav reposted

Excited to introduce DiffuseNNX, a comprehensive JAX/Flax NNX-based library for diffusion and flow matching! It supports multiple diffusion / flow-matching frameworks, Autoencoders, DiT variants, and sampling algorithms. Repo: github.com/willisma/diffu… Delve into details below!…


Sudhir Pratap Yadav reposted

It’s a bit easier to just explain how we got here. In the old days, memory and CPU ran at the same speed. The laws of physics make it incredibly hard to ramp up the MHz of the CPU while keeping physically separated memory chips in sync. The distance is just too far for the…

The question: "Why do CPUs have multiple cache levels?" often gets many CS students and professionals thinking and researching. This article from Fabian Giesen narrates a "cache story" in a relatable way, only to delve into the details - a must-read! fgiesen.wordpress.com/2016/08/07/why…

vivekgalatage's tweet image. The question: "Why do CPUs have multiple cache levels?" often gets many CS students and professionals thinking and researching.

This article from Fabian Giesen narrates a "cache story" in a relatable way, only to delve into the details - a must-read!

fgiesen.wordpress.com/2016/08/07/why…


Sudhir Pratap Yadav reposted

I really recommend watching this video on how to build high-performance motor controller. VERY interesting stuff!

VishakhRanotra's tweet image. I really recommend watching this video on how to build high-performance motor controller.

VERY interesting stuff!

Sudhir Pratap Yadav reposted

Have a look at kickstarter.com/projects/pcat2… this also does all the stuff u prolly need


Sudhir Pratap Yadav reposted

New Robotics Research from NVIDIA & Collaborators Exploring five innovative approaches in robot learning, manipulation, and simulation. @NVIDIARobotics


Sudhir Pratap Yadav reposted

< Choosing a Vision Backbone > your model’s backbone is its perspective pick ResNet, and it sees in edges pick a ViT, and it sees in patches the backbone decides how your model thinks here are some of the most practical backbones and when you should choose them, from the…

chesterzelaya's tweet image. &amp;lt; Choosing a Vision Backbone &amp;gt;

your model’s backbone is its perspective

pick ResNet, and it sees in edges
pick a ViT, and it sees in patches

the backbone decides how your model thinks

here are some of the most practical backbones and when you should choose them, from the…
chesterzelaya's tweet image. &amp;lt; Choosing a Vision Backbone &amp;gt;

your model’s backbone is its perspective

pick ResNet, and it sees in edges
pick a ViT, and it sees in patches

the backbone decides how your model thinks

here are some of the most practical backbones and when you should choose them, from the…

Sudhir Pratap Yadav reposted

Hoper's PCB has two distinct zones which are galvanically isolated using a wall of isolated DCDC converter, isolated transceivers and isolated drivers. There is no way for direct current to cross between zones. Power Zone: handles 60V battery input, motors power delivery, power…

observie's tweet image. Hoper&apos;s PCB has two distinct zones which are galvanically isolated using a wall of isolated DCDC converter, isolated transceivers and isolated drivers. There is no way for direct current to cross between zones.

Power Zone: handles 60V battery input, motors power delivery, power…

Sudhir Pratap Yadav reposted

A reminder that @corl_conf talks are fully and freely available to watch on @YouTube Keynotes are a must watch! Conference on Robot Learning youtube.com/@conferenceonr… #CoRL2025


Sudhir Pratap Yadav reposted

Below is mostly a summary of one of my favorite offline RL (ORL) reads: an amazing review by @nanjiang_cs and @tengyangx that clearly conveys what's hard and the key ideas used in that setting. nanjiang.cs.illinois.edu/files/STS_Spec… (screenshots are all from the pdf)

antoine_mln's tweet image. Below is mostly a summary of one of my favorite offline RL (ORL) reads: an amazing review by @nanjiang_cs and @tengyangx that clearly conveys what&apos;s hard and the key ideas used in that setting. nanjiang.cs.illinois.edu/files/STS_Spec… (screenshots are all from the pdf)

Our first #sim2real on single leg (of quadruped), policy trained in mujoco playground. We are currently 3 people working on rl + robotics. I will now start regularly posting updates of our various projects @vishcomestrue @aditya_kumar_bv


Loading...

Something went wrong.


Something went wrong.