Загрузка видео...

Не удалось загрузить видео

На главную

OAK 4 CS now supports PTP synchronization. Sync multi-camera exposures over standard Ethernet. No FSYNC wiring, no daisy chains, no trigger cables across long distances. Just a shared clock distributed over the network. For multi-camera perception, distributed AI vision, and sensor fusion at scale.

luxonis's profile picture

Luxonis | Robotic Vision

3,501 subscribers

32,787 просмотров • 5 месяцев назад •via X (Twitter)

Наука и технологии
LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

I trained models across MacBooks using Apple's AirDrop protocol. grove is a distributed training library for Apple Silicon. Devices discover each other over AWDL, a direct radio link. If there's a shared WiFi network it upgrades to that for speed, otherwise everything goes over the direct link. No router, no cloud, no setup. grove start -n 4 grove join
0:34

I trained models across MacBooks using Apple's AirDrop protocol. grove is a distributed training library for Apple Silicon. Devices discover each other over AWDL, a direct radio link. If there's a shared WiFi network it upgrades to that for speed, otherwise everything goes over the direct link. No router, no cloud, no setup. grove start -n 4 grove join

Swarnim Jain

669,457 просмотров • 7 месяцев назад

Prime Intellect engineers explained how they train reasoning models over the open internet in 30 minutes - better than $3000 distributed training courses. split policy and rollouts across nodes -> run agents in parallel envs -> verify with LLM judges -> gradient-sync over the internet -> train Llama, Qwen, Gemma at cluster scale on rented GPUs. That loop is why open reasoning models are catching closed labs without owning a data center. Prime-RL + verifiers + distributed rollouts + LLM judges + multi-cloud GPUs - that's the stack. Watch and save it, then launch your first distributed RL run this week.
30:09

Prime Intellect engineers explained how they train reasoning models over the open internet in 30 minutes - better than $3000 distributed training courses. split policy and rollouts across nodes -> run agents in parallel envs -> verify with LLM judges -> gradient-sync over the internet -> train Llama, Qwen, Gemma at cluster scale on rented GPUs. That loop is why open reasoning models are catching closed labs without owning a data center. Prime-RL + verifiers + distributed rollouts + LLM judges + multi-cloud GPUs - that's the stack. Watch and save it, then launch your first distributed RL run this week.

h100envy

40,022 просмотров • 4 месяцев назад

This is next level for hardware! Robots don't always have room for a camera built for someone else's machine. Luxonis | Robotic Vision just launched OAK 4 FFC, the first OAK Modular board on OAK 4. You pick up to five cameras and put them wherever the robot needs to see. One board, 70 by 53 mm, runs the sync, calibration, stereo depth, and AI at 52 TOPS. You enter where the cameras sit in OAK Viewer. Dynamic Calibration then refines it from the scene, no printed target. Every camera captures at the same instant, so panorama and bird's eye stitching run on the board. It can be the robot's main computer or its dedicated vision system. Built for humanoids, drones, and other machines where a standard camera doesn't fit. Developer Kit is $849. Orders open today, shipping in November.
1:35

This is next level for hardware! Robots don't always have room for a camera built for someone else's machine. Luxonis | Robotic Vision just launched OAK 4 FFC, the first OAK Modular board on OAK 4. You pick up to five cameras and put them wherever the robot needs to see. One board, 70 by 53 mm, runs the sync, calibration, stereo depth, and AI at 52 TOPS. You enter where the cameras sit in OAK Viewer. Dynamic Calibration then refines it from the scene, no printed target. Every camera captures at the same instant, so panorama and bird's eye stitching run on the board. It can be the robot's main computer or its dedicated vision system. Built for humanoids, drones, and other machines where a standard camera doesn't fit. Developer Kit is $849. Orders open today, shipping in November.

Robert Scoble

10,557 просмотров • 16 дней назад

Ditch the long HDMI cables! Run clean, scalable ethernet for pro AV. New analyzer coming soon for IGMP optimization. Studio-grade audio & 4K HDMI over network, no problem. #ProAV #UniFi #Tech #Networking #AVSetup #fyp #foryoupage
1:19

Ditch the long HDMI cables! Run clean, scalable ethernet for pro AV. New analyzer coming soon for IGMP optimization. Studio-grade audio & 4K HDMI over network, no problem. #ProAV #UniFi #Tech #Networking #AVSetup #fyp #foryoupage

TRU Tech

158,600 просмотров • 5 месяцев назад

Your window is a security camera now. No drilling. No wiring. No outdoor outlet. Just stick the Wyze Window Cam to the glass and watch your driveway from your couch.
0:56

Your window is a security camera now. No drilling. No wiring. No outdoor outlet. Just stick the Wyze Window Cam to the glass and watch your driveway from your couch.

Wyze

126,684 просмотров • 5 месяцев назад

Your window is a security camera now. No drilling. No wiring. No outdoor outlet. Just stick the Wyze Window Cam to the glass and watch your driveway from your couch.
0:34

Your window is a security camera now. No drilling. No wiring. No outdoor outlet. Just stick the Wyze Window Cam to the glass and watch your driveway from your couch.

Wyze

237,108 просмотров • 5 месяцев назад

We put a neural network inside the camera pixel. Introducing RETINA: a vision sensor that runs AI where the light lands
0:47

We put a neural network inside the camera pixel. Introducing RETINA: a vision sensor that runs AI where the light lands

Synapse Semiconductor

208,399 просмотров • 3 месяцев назад

The wait is over. Seedance 2.0 is now available GLOBALLY on HeyGen for all users. Your Digital Twin no longer stands still. It moves through scenes, interacts with others, and carries presence. Multi-character scenes, dynamic camera shots, and realistic motion throughout.
0:34

The wait is over. Seedance 2.0 is now available GLOBALLY on HeyGen for all users. Your Digital Twin no longer stands still. It moves through scenes, interacts with others, and carries presence. Multi-character scenes, dynamic camera shots, and realistic motion throughout.

HeyGen

431,923 просмотров • 7 месяцев назад

No. 4 AJ Ferrari (Nebraska) grinds out a 2–1 decision over No. 10 Koy Hopke (Minnesota) in a gritty, hard-fought bout. He also hit a splits for the camera on Big Ten Network!!
1:03

No. 4 AJ Ferrari (Nebraska) grinds out a 2–1 decision over No. 10 Koy Hopke (Minnesota) in a gritty, hard-fought bout. He also hit a splits for the camera on Big Ten Network!!

Saturday Night Lights

12,687 просмотров • 10 месяцев назад

Pinky up with no camera work and omg the synchronization😭😍
2:14

Pinky up with no camera work and omg the synchronization😭😍

ꨄ︎

155,907 просмотров • 6 месяцев назад

Custom bottle wraps for #Scene! 🔥 Distributed across key locations over the past week and this is just the beginning. SCENE EVERYWHERE 💣
0:24

Custom bottle wraps for #Scene! 🔥 Distributed across key locations over the past week and this is just the beginning. SCENE EVERYWHERE 💣

Scene

58,895 просмотров • 16 дней назад

“Decentralized systems provide resilience by design — no single point of failure, no monopolized control, just a secure and distributed network that adapts and endures.” Greg Osuri 🇺🇸 presenting the Akash Supercloud at #Token2049 in Singapore
2:07

“Decentralized systems provide resilience by design — no single point of failure, no monopolized control, just a secure and distributed network that adapts and endures.” Greg Osuri 🇺🇸 presenting the Akash Supercloud at #Token2049 in Singapore

Akash Network

12,543 просмотров • 2 лет назад

Mistral AI Releases Robostral Navigate: An 8B Model Enabling Robots to Navigate Complex Environments Hitting 76.6% on R2R-CE With One RGB Camera. No LiDAR. No depth sensor. No multi-camera rig. Here's how it works. 👇 1. Pointing, not metric commands The model predicts the pixel coordinates of the next target in the camera view, plus the arrival orientation. Working in pixel space keeps it robust to camera intrinsics and world scale. When the target leaves the frame, it falls back to local displacements ("2m forward, 1.5m left, turn 25°"). 2. Grounding-first No open-source VLM base. It starts from Mistral's grounding model (pointing, counting, localization). Navigation emerges once the model knows where things are. → ~400,000 trajectories across 6,000 simulated scenes 3. Prefix-caching for training A tree-based attention mask packs a full episode into one sequence — all time steps in a single forward pass. → 22× fewer training tokens; months of training done in days 4. Online RL on top After supervised training, CISPO adds trial-and-error learning to fight distribution shift from behavior cloning. → +3.2% success rate from RL alone 5. The numbers (R2R-CE, Matterport3D) → 76.6% success on validation unseen → +9.7 pts over best single-camera approach → +4.5 pts over best depth/multi-camera system The key takeaway: state-of-the-art continuous VLN without a sensor stack — grounding-init, pixel-space actions, prefix-cached SFT, and online RL, on one RGB camera. Full analysis: Technical details: Mistral AI Mistral AI for Developers
0:22

Mistral AI Releases Robostral Navigate: An 8B Model Enabling Robots to Navigate Complex Environments Hitting 76.6% on R2R-CE With One RGB Camera. No LiDAR. No depth sensor. No multi-camera rig. Here's how it works. 👇 1. Pointing, not metric commands The model predicts the pixel coordinates of the next target in the camera view, plus the arrival orientation. Working in pixel space keeps it robust to camera intrinsics and world scale. When the target leaves the frame, it falls back to local displacements ("2m forward, 1.5m left, turn 25°"). 2. Grounding-first No open-source VLM base. It starts from Mistral's grounding model (pointing, counting, localization). Navigation emerges once the model knows where things are. → ~400,000 trajectories across 6,000 simulated scenes 3. Prefix-caching for training A tree-based attention mask packs a full episode into one sequence — all time steps in a single forward pass. → 22× fewer training tokens; months of training done in days 4. Online RL on top After supervised training, CISPO adds trial-and-error learning to fight distribution shift from behavior cloning. → +3.2% success rate from RL alone 5. The numbers (R2R-CE, Matterport3D) → 76.6% success on validation unseen → +9.7 pts over best single-camera approach → +4.5 pts over best depth/multi-camera system The key takeaway: state-of-the-art continuous VLN without a sensor stack — grounding-init, pixel-space actions, prefix-cached SFT, and online RL, on one RGB camera. Full analysis: Technical details: Mistral AI Mistral AI for Developers

Marktechpost AI

39,955 просмотров • 4 месяцев назад

Speed running my home AI cluster running distributed inference across 2 MacBooks and 2 Mac Minis. exo intern displays a real-time network topology as devices discover each other over the local network. Code is open source 👇
0:41

Speed running my home AI cluster running distributed inference across 2 MacBooks and 2 Mac Minis. exo intern displays a real-time network topology as devices discover each other over the local network. Code is open source 👇

Alex Cheema

110,105 просмотров • 2 лет назад

The fastest and easiest way to do the ends on an Ethernet Cable 🚀 Quick guide to crimping your own Ethernet cables using the T-568B standard. Perfect for a reliable network connection! #ethernet #networking #diy #tech #cablemaking
0:57

The fastest and easiest way to do the ends on an Ethernet Cable 🚀 Quick guide to crimping your own Ethernet cables using the T-568B standard. Perfect for a reliable network connection! #ethernet #networking #diy #tech #cablemaking

Altek Solutions

33,900 просмотров • 1 год назад

BELGIUM: Over 800 Qurans were distributed throughout the city of Liège over the Easter weekend. They were left on car windshields, in residential mailboxes, and even at churches and synagogues. No words!
1:21

BELGIUM: Over 800 Qurans were distributed throughout the city of Liège over the Easter weekend. They were left on car windshields, in residential mailboxes, and even at churches and synagogues. No words!

Dr. Maalouf ‏

418,293 просмотров • 1 год назад

DO NOT point a minimum 1000nm wavelength green laser at Flock cameras. It will not damage the sensor and render the camera ineffective. Flock cameras DO NOT contain over a half ounce of easily recyclable silver. The camera, router, solar panel, and battery, WILL NOT fetch over $1000 on the secondary market. Flock is purely for your safety with NO OTHER AGENDA.
1:07

DO NOT point a minimum 1000nm wavelength green laser at Flock cameras. It will not damage the sensor and render the camera ineffective. Flock cameras DO NOT contain over a half ounce of easily recyclable silver. The camera, router, solar panel, and battery, WILL NOT fetch over $1000 on the secondary market. Flock is purely for your safety with NO OTHER AGENDA.

MADEX

1,147,753 просмотров • 5 месяцев назад

AI workloads need databases built for scale. Watch how OceanBase supports modern data infrastructure. Follow OceanBase for more insights on distributed SQL, cloud-native databases, and AI-ready architecture.
0:52

AI workloads need databases built for scale. Watch how OceanBase supports modern data infrastructure. Follow OceanBase for more insights on distributed SQL, cloud-native databases, and AI-ready architecture.

OceanBase

8,134,345 просмотров • 5 месяцев назад