Voxel51

Light mode in FiftyOne turns two! 🎉 Since we introduced it two years ago, we’ve seen all kinds of preferences for dark vs. light mode. So, we’re curious—which mode are you rocking in FiftyOne these days? #ComputerVision #OpenSource #MachineLearning #ML #AI #FiftyOne

London, here we come! Our AI, ML, and Computer Vision Meetup is happening on Nov. 19, and it’s shaping up to be a brilliant night of learning and connecting. Get a sneak peek at what we’ll be diving into: 🎙 Benchmarking and Optimization of LLMs with Ciera Fowler, ML Engineering Lead at Ori 🎙 Unlocking Value from Self-Driving Datasets with Daniel G., Machine Learning Engineer at Voxel51 🎙 Understanding Memory in AI Agents and Agentic Systems by Richmond Alake from MongoDB Don’t miss out! 📍 Fora Space Soho, 33 Broadwick St, London 🕒 5:30 PM - 8:30 PM ➡️ RSVP now: https://lnkd.in/gUub4R7G #AI #ML #MachineLearning #ComputerVision #MachineVision #DataScience

Ever wonder why your object detection model nails complex images but stumbles on the "easy" ones? 🤔 It’s not about your model architecture or training regime—it’s hidden patterns in your images that standard metrics miss. 💡 Think about factors like scene complexity, object-to-background ratios, and spatial distribution. These subtle, but powerful, elements shape your model’s performance more than you might expect. And by understanding and managing the data you're feeding your model, you can turn counterintuitive results into actionable insights. Want to dive deep and learn how to make sense of these patterns in your data? ➡️ Check out Harpreet Sahota 🥑's "Hands-On Data Centric AI" Coursera course here: https://lnkd.in/es6DjJms #AI #ML #ComputerVision #VisualAI #DataCuration

Data efficiency, state-of-the-art point tracking, AND faster performance? Sign us up! 🙋 Meta’s CoTracker3 is raising the bar in point tracking with: 🔥 Simplified architecture that keeps it lean and mean—CoTracker3 is now 27% faster than the previous fastest tracker 🔥 Data-efficient training that uses just 15,000 real videos to outperform models trained on millions 🔥 Enhanced real-world tracking capabilities, even for tricky occluded points 👉 Harpreet Sahota 🥑 dives into the most interesting improvements and innovations in CoTracker3 & more here: https://lnkd.in/eg-JjU-p #CoTracker3 #PointTracking #AI #ML #ComputerVision

Achtung! Join us on Nov 23 for the hands-on Berlin Visual AI Workshop focused on real-world computer vision challenges. Register 👉: https://lnkd.in/gpxch_6s The workshop is being hosted at MotionLab.Berlin and workshops will include: * Optimizing Berlin's Bike Routes: Hands-on Deep Learning with Street View Images - Antonio Rueda-Toicen at Hasso Plattner Institute * Building the Visual AI Datasets of Tomorrow: A Hands-on Workshop with Autonomous Vehicle Data - Daniel G. at Voxel51 #computervision #ai #artificialintelligence #machinelearning #datascience #machinevision #berlin #mobility #visualai

Join us on Dec 17 at the NexGen Cloud offices for the London AI, ML and Computer Vision Meetup. Reserve your spot: Talks will include: * The Frontier of Visual AI and Medical Imaging - Daniel G. at Voxel51 * The Hidden Accelerator: How Storage Can Elevate LLM Performance - Mischa van Kesteren at NexGen Cloud #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #gpu #nvidia Alessandro D'orsi

Voxel51 will be back in London on Nov 19 with another AI, Machine Learning and Computer Vision Meetup. Register: https://lnkd.in/gUub4R7G Join us at Fora Space in SOHO between 5:30 and 8:30 PM. * Understanding Memory in AI Agents and Agentic Systems - Richmond Alake at MongoDB * How to Unlock More Value from Self-Driving Datasets - Daniel G. at Voxel51 * Benchmarking and Optimization of LLMs - Ciera Fowler at Ori #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #gpu #mongodb

Join us on Nov 22 for the Berlin AI, ML and Computer Vision Meetup at MotionLab.Berlin. Reserve your spot: https://lnkd.in/g2avyUUX Talks will include: * Vector Streaming: The Memory Efficient Indexing for Vector Databases - Sonam Pankaj, Creator and Maintainer of Embed-Anything * When the Medium is the Message: Addressing Input Biases in Multimodal/Multilingual Models - Scott Martens at Jina AI * How to Unlock More Value from Self-Driving Datasets - Daniel G. at Machine Learning Engineer at Voxel51 #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #berlin

If you're into autonomous driving, this dataset is a treasure trove. 👇 Meet WayveScenes101—a next-level dataset that's loaded with synchronized, multi-camera scenes from diverse driving scenarios. Our ML Evangelist Harpreet Sahota 🥑 dives into the dataset, how to view it in FiftyOne format, and why it's a game-changer for anyone working in neural radiance fields or novel view synthesis. ⚠️ Warning: this dataset might just pull you in for hours—point clouds and all. #ComputerVision #AutonomousDriving #MachineLearning #VisualAI

Join us on Dec 12 for the AI, Machine Learning and Computer Vision Meetup! Register for the Zoom: https://lnkd.in/g7qrqyhg Talks and speakers will include: * How We Built CoTracker3: Simpler and Better Point Tracking by Pseudo-Labeling Real Videos - Nikita Karaev at Meta AI and University of Oxford * Hands-On with Meta AI's CoTracker3: Parsing and Visualizing Point Tracking Output - Harpreet Sahota 🥑 at Voxel51 #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #cotracker3 #metaai Andrea Vedaldi Christian Rupprecht