🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers Proposes SegFormer, a simple, efficient yet powerful semantic segmentation framework which unifies Transformers with MLP decoders. abs: arxiv.org/abs/2105.15203 code: github.com/NVlabs/SegForm…

Announcing SegFormer github.com/NVlabs/SegForm… arxiv.org/abs/2105.15203 SegFormer is a transformer based semantic segmentation framework with SOTA efficiency, accuracy and robustness. What's intriguing is the strong zero-shot robustness. Check out our demo: youtube.com/watch?v=J0MoRQ…

We observed that DNNs are not as lossy as expected during forward pass, despite generative and discriminative models. Hope to help the community understand DNNs deeper. With X. Dong Jose M. Alvarez Pavlo Molchanov Jan Kautz

1/2 papers accepted to ICCV2021 (and four others from the group). Biggest merits to Jiwoong, who is the first author. Work done while both me and him were interning with Jose M. Alvarez at NVIDIA AI. arxiv.org/abs/2103.16130

Panoptic SegFormer pdf: arxiv.org/pdf/2109.03814… abs: arxiv.org/abs/2109.03814

Expanding our Amazon Science team in Barcelona. Hiring IC scientists in computer vision, machine learning and NLP. Hiring at all levels. DM me if interested. #hiring Amazon Science #Barcelona Deep Learning Barcelona Symposium CVC_UAB Universitat Politècnica de Catalunya (UPC) IRI-Institut de Robòtica i Informàtica Industrial DataScienceUB ICREA Community

Happy to announce that our 𝐰𝐨𝐫𝐤𝐬𝐡𝐨𝐩 “Beyond Supervised Learning: Addressing Data Scarcity in Intelligent Transportation Systems” is happening on June 5th at #IV2022. Exciting speakers and a 𝐜𝐚𝐥𝐥 𝐟𝐨𝐫 𝐩𝐚𝐩𝐞𝐫𝐬: sites.google.com/view/iv2022-bs…. Kailun Yang IEEE Intelligent Transportation Systems Society

[#CVPR2022 oral] We introduce A-ViT, a method to enable Adaptive token computation for Vision Transformers. Let the network halt redundant tokens and compute only on active token deemed informative for the task, speeding up inference. Project: a-vit.github.io

A long overdue reply. Thank you AK ! In this work, we unveil the interesting trinity among grouping, IB and robustness in ViTs. We also propose FAN, a general-purpose ViT backbone with SOTA robustness. 📢Check our work at #ICML2022 Github: github.com/NVlabs/FAN 🧵

87% of AI projects fail to hit production, 96% have data problems & 51% underestimate how much data they need. Our #NeurIPS2022 paper proposes learning & optimizing data collection! Sanja Fidler Rafid Mahmood James Lucas and Marc-Law lnkd.in/gmqiyQah

Looking for faster and more accurate models to be deployed on the edge? Our #eccv2022 paper on channel pruning focuses on improving the latency-accuracy trade-off for large pruning ratios. Code: lnkd.in/gJMctMpc Paper: lnkd.in/gX3S8Z9w In…lnkd.in/gHXwsbmH

Looking for faster and more accurate models to be deployed on the edge? Our #eccv2022 paper on channel pruning focuses on improving the latency-accuracy trade-off for large pruning ratios. Code: lnkd.in/gJMctMpc Paper: lnkd.in/gX3S8Z9w In…lnkd.in/gbA352EZ

Excited to share our 2 papers to be presented at #NeurIPS2022 this week in New Orleans, USA. 1. Optimizing Data Collection for Machine Learning Paper Pdf: arxiv.org/abs/2210.01234 2. Structural Pruning via Latency-Saliency Knapsack Paper Pdf: arxiv.org/abs/2210.06659

We are looking for highly motivated research and SWE interns (all year round) to work with us on Perception for Autonomous Vehicles at Nvidia. We are particularly interested in efficient deep learning, reducing labeling costs, be…lnkd.in/gFr_aR4j lnkd.in/gsZf84-P

Our Autonomous Vehicle Perception Research team at #nvidia has several year-round openings for strong research #intern (PhD and MS, preferably). Ideally, a candidate with proven track record: research publications in top-tier venues such as #neurips , #i…lnkd.in/gbxrUN7x

Workshop on Autonomous Driving #CVPR2025 Paper submission is now open! cvpr2023.wad.vision Paper submission deadline: *6* March 2023 (23:59 PST) Notification to authors: 31 March 2023

lnkd.in/gcsnNnWA

#NVIDIAResearch wins #CVPR2023 #autonomousdriving challenge and #innovation award, in advance of conference opening Sunday in Vancouver. Learn more in our blog #automotive #AI nvda.ws/465JJ2G

Wanna learn the best ways to train and deploy your large neural networks on GPU? We will host the "Full-stack GPU-based acceleration of deep learning" tutorial at #cvpr2023 tomorrow June 19th at east Hall 11, 1:30pm to 5:00pm. Join…lnkd.in/gkkpyyU9 lnkd.in/gmDQCMTV

Slides for the "Full-stack GPU-based acceleration of deep learning" tutorial at #cvpr2023 are now available. 1. Foundations of Deep Learning hardware lnkd.in/ggdJQ7zU 2. DNN Performance optimization: How to achieve more w…lnkd.in/grvHa54w lnkd.in/gmDQCMTV