Udit Singh Parihar

MS in CSE with specialization in Robotics at IIITH, working on SLAM, Computer Vision and Deep Learning.

[Abhishek Gupta @abhishekunique7]

Constructing interactive simulated worlds has been a challenging problem, requiring considerable manual effort for asset creation and articulation, and composing assets to form full scenes. In our new work - DRAWER, we made the process of creating scenes in simulation as simple as taking a video of the scene and out comes a high-quality, fully interactive environment in simulation. No human simulation designer involved!

https://t.co/JVA5Nap2fe

A 🧵(1/7)

[Jack Morris @jxmnop]

posted the other day about model distillation. pretty much everyone responded with their theories

professors, leading lab researchers, students, pseudoanonymous anime-profile posters

seems there's no clear consensus why it works, but here are the theories 🧵

[Chen-Hsuan Lin @chenhsuanlin]

🌌 NVIDIA Cosmos -- our World Foundation Model platform! Super excited to have made core contributions in multiple aspects. Physical AI is key to modeling the universe of worlds 🌎!

75-page tech report 📄: https://t.co/eUB816Fhzz
Try them out now 😲! https://t.co/7wKY3Tm006

[Simon Boehm @Si_Boehm]

I wrote the most naive CUDA matrix multiply and iteratively optimised it to ~80% of cuBLAS performance:
https://t.co/zHBigTKhCB

[Shubodh Sai 🤖 @shubodhs_ai]

🚗 LIP-Loc: LiDAR Image Pretraining for Cross-Modal Localization

LiDAR Image Pretraining meets Visual Localization🤝

tl;dr: CLIP but on domains of LiDAR and images for localisation

Project page: https://t.co/nN7HAPMhbF
Code: https://t.co/P3EbrOLSbC

[Shubham Tulsiani @shubhtuls]

[1/6] What representation comes to mind when you think of a ‘camera’? Perhaps an extrinsic + intrinsic matrix? In our ICLR (oral) paper, we instead infer a distributed representation where each pixel is associated with a ray, and show SoTA results for few-view pose estimation.

[Udit Singh Parihar @udit_parihar]

@ShantanuSingh91 I am at Bangalore airport and I have been waiting since afternoon and sleeping at the gate. My friend was ahead in and got boarded and I am still waiting to board the flight. Very bad service from Indigo

#recession #KisanAndolan2024 #NVIDIA #Indigo

[Jayaram Reddy @Jayaram_Reddy10]

(1/7) 🧵Excited to share our latest work : EDMP (Ensemble-of-costs-guided Diffusion for Motion Planning)

Bridging Classical and Deep Learning based methods in Robotic Motion Planning.

Abstract: https://t.co/dB1t5pXVnO
Project Page: https://t.co/YxZVniM3cI

[Yi Tay @YiTayML]

We’re coming out of stealth with $58M in funding to build generative models and advance AI research at @RekaAILabs 🔥🚀

Language models and their multimodal counterparts are already ubiquitous and massively impactful everywhere.

That said, we are still at the beginning of this new revolution. Over the next decade and beyond, it is inevitable that we will witness significant innovation in the field as more capabilities are discovered and breakthroughs are being made.

We founded @RekaAILabs to be at the helm of innovation.

We have two key goals - to build amazing generative models and to push the frontiers of AI research.

Check out more details in our launch announcement below 👇

[Zhengqi Li @zhengqi_li]

Check out our CVPR 2023 Award Candidate paper, DynIBaR! https://t.co/8IF8GWXKHR
DynIBaR takes monocular videos of dynamic scenes and renders novel views in space and time. It addresses limitations of prior dynamic NeRF methods, rendering much higher quality views.

[Udit Singh Parihar @udit_parihar]

@gmiller @Aella_Girl What if she referring to units in nanometers

[Udit Singh Parihar @udit_parihar]

@karenxcheng @Adobe “Compensation structure for artists whose art was being used for training image generation model”
By that logic openai, google, meta should also pay to people whose code/blogs are used during training of gpt/llms.

[Udit Singh Parihar @udit_parihar]

@GaryMarcus @DrJimFan I don’t understand the open book part, since gpt 4 is not browsing the internet and just giving answers from its memory (weights) then how it it open book?

[Udit Singh Parihar @udit_parihar]

@simonw Can we really say that 13B llama.cpp is comparable to GPT 3. Since in llama.cpp you are doing aggressive quantisation. Have someone compared how these new quantised weights performed compared to gpt 3?

[Pete Florence @peteflorence]

Today we share more on PaLM-E! (https://t.co/7fJdwhFTg2)
Thread 🧵with blog post link at the end.

PaLM-E can do a lot of things across robotics, vision, and language… but let’s look at a few capabilities in detail, step by step 😉
👇

[Cameron R. Wolfe, Ph.D. @cwolferesearch]

Have you heard of Contrastive Language Image Pre-training (CLIP)? It's a cutting-edge deep learning technique that learns a joint embedding space between images and text by training over tons of image-caption pairs from the internet... 🧵[1/7]

[Adam D'Angelo @adamdangelo]

We are excited to launch Poe for desktop today at https://t.co/cMuZ6Cb3cY. With this launch, between our existing iOS app this new web interface, we believe Poe is now the fastest and easiest way to talk to AI across iOS and all desktop platforms.

[Delip Rao e/σ @deliprao]

Why should I pay $20 a month for ChatGPT Pro when I have a cheap API alternative? Anyone has ideas?

[Udit Singh Parihar @udit_parihar]

@code Install copilot

[Udit Singh Parihar @udit_parihar]

@chheplo @julien_c Is that the reason they put double L in Llama, twice failed?