Yichen Sheng
@Coding_Black
Research scientist in NVIDIA. Working in graphics and vision. Opinions are my own.
You might like
So happy to see #DL3DV-10K boradly used. People asked me the born of #DL3DV. DL3DV was a student-led dataset without any funding from professors or any company. Every data collection team member worked really hard. Proud of the team and can't wait to see the future!
Thanks @_akhaliq. It takes a lot of effort to collect this large-scale video dataset (more than 10K video) for 3D vision community. We hope you enjoy the gift!
Heading to #ICCV2025 ? Do not miss tomorrow's DLSS4 keynote talk by @edliu1105 Time: 1:30~2:15 pm, 20th Oct ROOM: B317
🚀🚀🚀
The 3 Coolest Things I saw at #Siggraph this year: 1. Nvidia's DLSS Denoiser in Blender This makes Cycles basically real-time. Meaning 3d artists can create with almost zero latency. Not yet available to the public, but it's expected to make it into a future Blender release.
One of our roles in LLM/VLM research at NVIDIA is to explore effective data recipes for training large-scale models and share them to the public—an area where transparency has been limited, as seen with models like Gemini, GPT-4o, Qwen-VL models etc. The Eagle2 project aligns…
The Eagle 2 paper from Nvidia is such a goldmine.
@CVPR2025 Wei Xiong and I will host a coffee chat session(11. 30～12. 30) in Room211 for Nvidia university hiring. Feel free to come and say hi if you are interested in DLSS and next generation generative rendering. linkedin.com/posts/nvidia-u…
Congrats to "VGGT" wins the best paper award at #CVPR2025 ! We are happy that #DL3DV benefits "VGGT" and the community. We will host the #DL3DV Demo session this afternoon from 4:00-6:00 pm. Come by and see what is the new in DL3DV!
Many Congratulations to @jianyuan_wang, @MinghaoChen23, @n_karaev, Andrea Vedaldi, Christian Rupprecht and @davnov134 for winning the Best Paper Award @CVPR for "VGGT: Visual Geometry Grounded Transformer" 🥇🎉 🙌🙌 #CVPR2025!!!!!!
🚀 We just open-sourced Cosmos DiffusionRenderer! This major upgrade brings significantly improved video de-lighting and re-lighting—powered by NVIDIA Cosmos and enhanced data curation. Released under Apache 2.0 and Open Model License. Try it out! 🔗 github.com/nv-tlabs/cosmo…
🚀 Introducing DiffusionRenderer, a neural rendering engine powered by video diffusion models. 🎥 Estimates high-quality geometry and materials from videos, synthesizes photorealistic light transport, enables relighting and material editing with realistic shadows and reflections
Vision Language Models can be amazing at document understanding. Please check out our Nano sized model. More to come!
🥇Our NVIDIA Llama Nemotron Nano VL model is #1 on the OCRBench V2 leaderboard. Designed for advanced intelligent document processing and understanding, this model extracts diverse info from complex documents with precision, all on a single GPU. 📗 Get the technical details…
Generative photography is our new physics-based camera-aware image generation tool. To appear @CVPR 2025 (Highlight paper 13.5%) arxiv.org/abs/2412.02168 - Respect camera controls - Photorealistic generation Credits to Yu Yuan @PurdueECE
Introducing AceMath-RL-Nemotron-7B, an open math model trained with reinforcement learning from the SFT-only checkpoint: Deepseek-R1-Distilled-Qwen-7B. It achieves: - AIME24: 69.0 (+13.5 gain by RL) - AIME25: 53.6 (+14.4) - LiveCodeBench: 44.4 (surprisingly, +6.8 gain after…
Eagle2.5 natively supports long-context without using any compression module. Eagle2.5-8B has: • got 6 out of 10 SOTA on long video benchmarks • beat GPT-4o (0806) on 3/5 video tasks • beat Gemini 1.5 Pro on 4/6 video tasks • got SOTA result on Hour-long video benchmark.
A long time ago, back before DLSS was in many games (and when my hair was shorter and less gray), I went to Nintendo HQ to show them an early prototype of DLSS 2, in the hopes that a future Switch console would use DLSS. I'm so proud that the Switch 2 will be DLSS powered!…
Even though we haven't been publishing papers around DLSS, it takes a tremendous amount of hardcore research to bring AI models like DLSS4 into production. Super happy that we're sharing technical insights on DLSS4 in this report. So proud of the teams!
Technical insights into what makes DLSS4 so great, including MFG, transformer models, and Reflex Frame Warp: research.nvidia.com/labs/adlr/DLSS…
Tired of slow diffusion models? Our new paper introduces f-distill, enabling arbitrary f-divergence for one-step diffusion distillation. JS divergence gives SOTA results on text-to-image! Choose the divergence that suits your needs. Joint work with @wn8_nie @ArashVahdat 1/N
github.com/NVIDIA/Cosmos Cosmos is a developer-first platform designed to help physical AI builders accelerate their development. It has pre-trained world foundation models (diffusion & autoregressive) in different sizes and video tokenizers. They are open models with permissive…
Big Congratulations to our team! We firmly believe AI will shape the future of the graphics. If you are also passionate and want to make a history, please join us!
DLSS 4 is here!🚀Multi-Frame Generation for smoother frame rates and Transformer-based models driving enhanced image quality. This milestone reflects the passion, creativity, and dedication of many people at NVIDIA. Excited to shape the future of graphics with AI? Join us!
Our team at NVIDIA is continuously looking for highly motivated interns to work on intelligence in audio understanding and synthesis. Please reach out if you would like to collaborate with us!
We equip the diffusion model with a camera lens feature. You will feel safe when you *just* want to edit the focal length/shutter/bokeh/color temperature as the contents are preserved . Arxiv: arxiv.org/abs/2412.02168
NVIDIA has found a way to add camera physics to diffusion models. Literally makes it possible to generate consistent images but with a different aperture, focal length, shutter speed or color temperature.
📢Please check out our newest work on feed-forward reconstruction of dynamic monocular videos! With our bullet-time formulation, we reach great flexibility and state-of-the-art performance!
Excited to introduce BTimer: a real-time solution for reconstructing dynamic scenes from monocular videos! 🚀 Generate per-frame (bullet-time) 3D GS scenes in 150ms on a single GPU, achieving quality that is on par with the optimization methods! Project: research.nvidia.com/labs/toronto-a…
United States Trends
- 1. #FinallyOverIt 5,622 posts
- 2. Summer Walker 17.7K posts
- 3. 5sos 22.1K posts
- 4. #criticalrolespoilers 4,158 posts
- 5. Justin Fields 10.1K posts
- 6. #GOAI 3,165 posts
- 7. #zzzSpecialProgram 2,939 posts
- 8. Wale 33.8K posts
- 9. #MakeOffer 12.5K posts
- 10. Go Girl 26.3K posts
- 11. Bihar 262K posts
- 12. 1-800 Heartbreak 1,461 posts
- 13. Domain For Sale 13K posts
- 14. Drake Maye 21.4K posts
- 15. Judge 207K posts
- 16. Robbed You 4,161 posts
- 17. Jalen Johnson 9,054 posts
- 18. Disc 2 1,025 posts
- 19. RIN AOKBAB BEGIN AGAIN 501K posts
- 20. Alignerz 155K posts
You might like
-
Simon Giebenhain
@SGiebenhain -
Xiuming Zhang
@xiuming_zhang -
He Zhang
@zhanghesprinter -
Cheng Lin
@_cheng_lin -
Hao Tang
@HaoTang_ai -
Zhiyang (Frank) Dou
@frankzydou -
Linyi Jin
@jin_linyi -
Jingbo Wang
@Alex_wangjingbo -
Yiming Xie
@YimingXie4 -
Giljoo Nam
@giljoo_nam -
Yao (Ava) Xiao
@AvaXYao -
Jiawei Yang
@JiaweiYang118 -
Subhransu Maji
@MajiSubhransu -
Yueh-Cheng Liu
@liuyuehcheng -
Guangyu Shen
@guangyuNoah
Something went wrong.
Something went wrong.