Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

Introducing Distil-Whisper v3 ⚡ > ~50% less parameters and 6x faster than Large-v3. > More accurate than large-v3 on long-form synthesis. Available with 🦀 WebGPU, Whisper.cpp, Transformers, Faster-Whisper and Transformers.js support! Drop in; no changes are required! 🔥

reach_vb's profile picture

Vaibhav (VB) Srivastav

43,449 subscribers

90,651 Aufrufe • vor 2 Jahren •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

Released initial support for WebGPU/TSL in @takram/three-atmosphere. It features much faster LUT generation, physically correct lighting at large-scale scenes, geometry-based lens flare/glare effect and more. It took longer than I anticipated.
0:18

Released initial support for WebGPU/TSL in @takram/three-atmosphere. It features much faster LUT generation, physically correct lighting at large-scale scenes, geometry-based lens flare/glare effect and more. It took longer than I anticipated.

Shota Matsuda 松田 聖大

19,856 Aufrufe • vor 8 Monaten

Day 17 of why you need to use Jupiter Mobile. FAST AS F*** Hate getting nuked on entries and exits? Slower execution means more slippage and worse prices. Every swap on Jupiter Mobile runs on Ultra V3, with sub‑second precision, 0–1 block execution, and 3x more trades landed, all while paying ~10x lower fees than other wallets. The team has been obsessing over latency, routing, and execution, so your trades hit faster and cleaner, every time. Just Use Jupiter (Mobile).
0:10

Day 17 of why you need to use Jupiter Mobile. FAST AS F*** Hate getting nuked on entries and exits? Slower execution means more slippage and worse prices. Every swap on Jupiter Mobile runs on Ultra V3, with sub‑second precision, 0–1 block execution, and 3x more trades landed, all while paying ~10x lower fees than other wallets. The team has been obsessing over latency, routing, and execution, so your trades hit faster and cleaner, every time. Just Use Jupiter (Mobile).

Jupiter

177,278 Aufrufe • vor 5 Monaten

NO MORE FAILED TRANSACTIONS. Releasing BOLT PRO, a massive upgrade to our BOLT feature now available for 10,000 new users. With BOLT PRO swaps (manual, limit orders and DCA): - take less than 2 seconds to go through on average. - almost never fail or timeout. The biggest edge you can have in this market is speed, and with BOLT you are guaranteed to be faster than all other traders. 🐴🤖
0:16

NO MORE FAILED TRANSACTIONS. Releasing BOLT PRO, a massive upgrade to our BOLT feature now available for 10,000 new users. With BOLT PRO swaps (manual, limit orders and DCA): - take less than 2 seconds to go through on average. - almost never fail or timeout. The biggest edge you can have in this market is speed, and with BOLT you are guaranteed to be faster than all other traders. 🐴🤖

Trojan Trading

63,186 Aufrufe • vor 2 Jahren

BREAKING: Tesla has just officially announced their next-gen V4 Supercharger cabinet, capable of 500kW charging speeds for cars and 1.2MW for Semi. HOLY SHIT! Highlights: • Faster charging: Supports 400V-1000V vehicle architectures, including 30% faster charging for Cybertruck. S3XY vehicles enjoy 250kW charge rates they already experience on V3 Cabinet — charging up to 200 miles in 15 minutes. • Faster deployments: V4 Cabinet powers 8 posts, 2X the stalls per cabinet. Lower footprint and complexity = more sites coming online faster. • Next-generation hardware: Cutting-edge power electronics designed to be the most reliable on the planet, with 3X power density enabling higher throughput with lower costs. First sites with V4 Cabinets are going into permitting now. First openings in 2025.
0:19

BREAKING: Tesla has just officially announced their next-gen V4 Supercharger cabinet, capable of 500kW charging speeds for cars and 1.2MW for Semi. HOLY SHIT! Highlights: • Faster charging: Supports 400V-1000V vehicle architectures, including 30% faster charging for Cybertruck. S3XY vehicles enjoy 250kW charge rates they already experience on V3 Cabinet — charging up to 200 miles in 15 minutes. • Faster deployments: V4 Cabinet powers 8 posts, 2X the stalls per cabinet. Lower footprint and complexity = more sites coming online faster. • Next-generation hardware: Cutting-edge power electronics designed to be the most reliable on the planet, with 3X power density enabling higher throughput with lower costs. First sites with V4 Cabinets are going into permitting now. First openings in 2025.

Sawyer Merritt

1,196,510 Aufrufe • vor 1 Jahr

Our newest AI accelerator Maia 200 is now online in Azure. Designed for industry-leading inference efficiency, it delivers 30% better performance per dollar than current systems. And with 10+ PFLOPS FP4 throughput, ~5 PFLOPS FP8, and 216GB HBM3e with 7TB/s of memory bandwidth it's optimized for large-scale AI workloads. It joins our broader portfolio of CPUs, GPUs, and custom accelerators, giving customers more options to run advanced AI workloads faster and more cost-effectively on Azure.
0:11

Our newest AI accelerator Maia 200 is now online in Azure. Designed for industry-leading inference efficiency, it delivers 30% better performance per dollar than current systems. And with 10+ PFLOPS FP4 throughput, ~5 PFLOPS FP8, and 216GB HBM3e with 7TB/s of memory bandwidth it's optimized for large-scale AI workloads. It joins our broader portfolio of CPUs, GPUs, and custom accelerators, giving customers more options to run advanced AI workloads faster and more cost-effectively on Azure.

Satya Nadella

847,880 Aufrufe • vor 5 Monaten

We’re introducing Cue-Pro — a new way to visualize and execute edit predictions, built to keep you in flow. Instead of isolated edit points, Cue-Pro brings repository-level edit predictions into your IDE, with a side panel that provides a bird’s-eye view of what needs to change. Same workflow: tab to edit, tab to jump. Better outcomes: - Faster, more consistent edits - Fewer missed changes across large codebases - Repository-wide edit predictions that are visual and navigable — without breaking flow In beta testing, Cue-Pro delivered in +19% code acceptance rate and +20% improvement in characters per opportunity (CPO). (Below are a few common scenarios where Cue-Pro helps you ship faster ↓)
0:11

We’re introducing Cue-Pro — a new way to visualize and execute edit predictions, built to keep you in flow. Instead of isolated edit points, Cue-Pro brings repository-level edit predictions into your IDE, with a side panel that provides a bird’s-eye view of what needs to change. Same workflow: tab to edit, tab to jump. Better outcomes: - Faster, more consistent edits - Fewer missed changes across large codebases - Repository-wide edit predictions that are visual and navigable — without breaking flow In beta testing, Cue-Pro delivered in +19% code acceptance rate and +20% improvement in characters per opportunity (CPO). (Below are a few common scenarios where Cue-Pro helps you ship faster ↓)

TRAE

453,248 Aufrufe • vor 6 Monaten

NEW STARLINK GIGABIT DISHES INCOMING—FASTER THAN FIBER? SpaceX is turning up the heat with its next-gen Starlink gigabit connection gear! New dishes are in the works: * 2 Gigs, No Wires – New dish pushes speeds up to 2 Gbps. Blink and you’ll miss your download. * Next-Level Mini – A souped-up follow-up to the Starlink Mini is already in the pipeline. * Waiting on Launch – Needs FCC clearance and a Starship boost to get those V3 satellites in orbit. * Fiber, Who? – Tighter beams. Lower lag. Bigger bandwidth. Space-based internet just got real. Expect business users to get first dibs for now—but don’t blink! The future of lightning-fast satellite internet is closing in. Source: PCMag
0:17

NEW STARLINK GIGABIT DISHES INCOMING—FASTER THAN FIBER? SpaceX is turning up the heat with its next-gen Starlink gigabit connection gear! New dishes are in the works: * 2 Gigs, No Wires – New dish pushes speeds up to 2 Gbps. Blink and you’ll miss your download. * Next-Level Mini – A souped-up follow-up to the Starlink Mini is already in the pipeline. * Waiting on Launch – Needs FCC clearance and a Starship boost to get those V3 satellites in orbit. * Fiber, Who? – Tighter beams. Lower lag. Bigger bandwidth. Space-based internet just got real. Expect business users to get first dibs for now—but don’t blink! The future of lightning-fast satellite internet is closing in. Source: PCMag

Mario Nawfal

684,310 Aufrufe • vor 1 Jahr

Today, we’ve successfully launched Band Oracle v3 Testnet Phase 2 — our biggest upgrade yet. This release makes BandChain faster, more transparent, and more interoperable than ever. Let’s break down what’s new👇 What did we achieve? — 3x Faster blocktime from 3s —> 1s — 10x symbols cap expand from 100 —> 300 symbols — 10x Higher Throughputs: 40k to 400k TX/day — 3x Max TXs Support/ Day: 7.2M TXs —> 21.6M TXs 🌐 New Data Tunnel routes are LIVE We’re expanding Band’s oracle reach across the multi-chain world: —> Cosmos - The Interchain ⚛️’s IBC-Hook: seamless interchain contract querying —> Router Protocol: Bridging & EVM chains & Soon on Solana —> Axelar Network: ongoing collab for cross-chain verified feeds Interoperability is no longer a wishlist — it’s real. 🛠️ Minor upgrades with major impact: — Cylinder CLI runs standalone — Signal times are now block-based (not machine time) — Auto-select TSS groups — Yoda now auto-bumps gas to avoid TX failure — Telemetry for BeginBlock/EndBlock durations It’s all about smoother ops for devs and validators. 🧪 What’s next? ✅ Recap + docs for Testnet Phase 2 ✅ More Data Tunnel stress testing ✅ Backend + validator-side optimizations 🚀 Mainnet Launch in Q3 2025 We’re almost there. Band Oracle v3 is becoming the go-to open data oracle for all ecosystems. Fast. Transparent. New standard for Multi-chain native. Thanks to the validators, devs, partners, and Band community for building this with us. Let’s keep going. Read the full announcement here:
0:10

Today, we’ve successfully launched Band Oracle v3 Testnet Phase 2 — our biggest upgrade yet. This release makes BandChain faster, more transparent, and more interoperable than ever. Let’s break down what’s new👇 What did we achieve? — 3x Faster blocktime from 3s —> 1s — 10x symbols cap expand from 100 —> 300 symbols — 10x Higher Throughputs: 40k to 400k TX/day — 3x Max TXs Support/ Day: 7.2M TXs —> 21.6M TXs 🌐 New Data Tunnel routes are LIVE We’re expanding Band’s oracle reach across the multi-chain world: —> Cosmos - The Interchain ⚛️’s IBC-Hook: seamless interchain contract querying —> Router Protocol: Bridging & EVM chains & Soon on Solana —> Axelar Network: ongoing collab for cross-chain verified feeds Interoperability is no longer a wishlist — it’s real. 🛠️ Minor upgrades with major impact: — Cylinder CLI runs standalone — Signal times are now block-based (not machine time) — Auto-select TSS groups — Yoda now auto-bumps gas to avoid TX failure — Telemetry for BeginBlock/EndBlock durations It’s all about smoother ops for devs and validators. 🧪 What’s next? ✅ Recap + docs for Testnet Phase 2 ✅ More Data Tunnel stress testing ✅ Backend + validator-side optimizations 🚀 Mainnet Launch in Q3 2025 We’re almost there. Band Oracle v3 is becoming the go-to open data oracle for all ecosystems. Fast. Transparent. New standard for Multi-chain native. Thanks to the validators, devs, partners, and Band community for building this with us. Let’s keep going. Read the full announcement here:

Band

17,921 Aufrufe • vor 1 Jahr

✨ Every time the video models get better, the try on model on Photo AI also becomes a lot more useful, as a large % of my customers now are e-commerce store And showing clothes in a video is nice for sales! With AI this means stores don't need to do expensive shoots flying a model and entire camera and light crew around the world They can just upload a few photos of their models, then upload the clothes, and describe the setting (like a beach in Thailand) and in less than 10 seconds it's generated, for a video in less than a minute! Below is the input: a dress laid flat, and output: a full video shoot
0:10

✨ Every time the video models get better, the try on model on Photo AI also becomes a lot more useful, as a large % of my customers now are e-commerce store And showing clothes in a video is nice for sales! With AI this means stores don't need to do expensive shoots flying a model and entire camera and light crew around the world They can just upload a few photos of their models, then upload the clothes, and describe the setting (like a beach in Thailand) and in less than 10 seconds it's generated, for a video in less than a minute! Below is the input: a dress laid flat, and output: a full video shoot

@levelsio

334,170 Aufrufe • vor 1 Jahr

Self-Calibrating Gaussian Splatting for Large Field of View Reconstruction Note: Check below for full video. Abstract (cited): "In this paper, we present a self-calibrating framework that jointly optimizes camera parameters, lens distortion, and 3D Gaussian representations, enabling accurate and efficient scene reconstruction. Our technique is particularly effective for high-quality scene reconstruction from large field-of-view (FOV) imagery taken with wide-angle lenses, allowing the scene to be modeled from a smaller number of images. We introduce a novel method for modeling complex lens distortions using a hybrid network that combines invertible residual networks with explicit grids. This design effectively regularizes the optimization process, achieving greater accuracy than conventional camera models. Additionally, we propose a cubemap-based resampling strategy to support large FOV images without sacrificing resolution or introducing distortion artifacts. Our method is compatible with the fast rasterization of Gaussian Splatting, adaptable to a wide variety of camera lens distortions, and demonstrates state-of-the-art performance on both synthetic and real-world datasets."
0:16

Self-Calibrating Gaussian Splatting for Large Field of View Reconstruction Note: Check below for full video. Abstract (cited): "In this paper, we present a self-calibrating framework that jointly optimizes camera parameters, lens distortion, and 3D Gaussian representations, enabling accurate and efficient scene reconstruction. Our technique is particularly effective for high-quality scene reconstruction from large field-of-view (FOV) imagery taken with wide-angle lenses, allowing the scene to be modeled from a smaller number of images. We introduce a novel method for modeling complex lens distortions using a hybrid network that combines invertible residual networks with explicit grids. This design effectively regularizes the optimization process, achieving greater accuracy than conventional camera models. Additionally, we propose a cubemap-based resampling strategy to support large FOV images without sacrificing resolution or introducing distortion artifacts. Our method is compatible with the fast rasterization of Gaussian Splatting, adaptable to a wide variety of camera lens distortions, and demonstrates state-of-the-art performance on both synthetic and real-world datasets."

MrNeRF

17,206 Aufrufe • vor 1 Jahr

You are probably breathing TOO MUCH, which is affecting your longevity and looks. It probably sounds counterintuitive - more breaths = better, right? No. Today, we breathe shallow, from our mouths, into our chests at a pace that is more to our detriment than otherwise. The average person takes 18-20 breaths per minute, optimal pace being 6-10. This blows out CO2 faster than the body can produce it. Less CO2 means less oxygen reaches your cells, no matter how much air you breathe in. When you chronically overbreathe, your: → Sleep collapses → Skin dulls and ages faster → Cortisol stays elevated 24/7 → Lymph backs up and the face puffs → Lips lose tone and the seal weakens → Forward head posture sets in and the jaw retracts → Tongue drops from the palate and the maxilla narrows → Cheekbones lose support and the midface collapses inward Just from breathing incorrectly. The most ideal way to breathe is: -through your nose -funnelling the air BACKWARDS (imagine directing the air into the lower back region) -in a slowly and controlled manner This is called Back Breathing and you will instantly feel a difference. Breathe less. Live longer. Look better. ⚡️
0:16

You are probably breathing TOO MUCH, which is affecting your longevity and looks. It probably sounds counterintuitive - more breaths = better, right? No. Today, we breathe shallow, from our mouths, into our chests at a pace that is more to our detriment than otherwise. The average person takes 18-20 breaths per minute, optimal pace being 6-10. This blows out CO2 faster than the body can produce it. Less CO2 means less oxygen reaches your cells, no matter how much air you breathe in. When you chronically overbreathe, your: → Sleep collapses → Skin dulls and ages faster → Cortisol stays elevated 24/7 → Lymph backs up and the face puffs → Lips lose tone and the seal weakens → Forward head posture sets in and the jaw retracts → Tongue drops from the palate and the maxilla narrows → Cheekbones lose support and the midface collapses inward Just from breathing incorrectly. The most ideal way to breathe is: -through your nose -funnelling the air BACKWARDS (imagine directing the air into the lower back region) -in a slowly and controlled manner This is called Back Breathing and you will instantly feel a difference. Breathe less. Live longer. Look better. ⚡️

Oscar Patel

25,317 Aufrufe • vor 1 Monat

🆕 CrowdStrike is acquiring Onum to supercharge autonomous cybersecurity with real-time data pipelines. If Falcon SIEM is the engine of the modern SOC, Onum is both the pipeline + filter — streaming high-quality, filtered fuel quickly into the engine to drive more efficient and superior performance. Onum delivers transformational advantages across three critical dimensions: ⚡ Speed: Delivers 5X more events per second than its nearest competitor and processes security and observability data in real-time versus legacy batch and store methods. ⚡ Cost: Smart filtering reduces data storage costs by 50% through intelligent optimization. ⚡ Superior Outcomes: Real-time pipeline detection starts before data enters the Falcon platform, delivering up to 70% faster incident response with 40% less ingestion overhead. The future of cybersecurity is here – built on real-time, high-fidelity data. 👉 Learn more:
0:10

🆕 CrowdStrike is acquiring Onum to supercharge autonomous cybersecurity with real-time data pipelines. If Falcon SIEM is the engine of the modern SOC, Onum is both the pipeline + filter — streaming high-quality, filtered fuel quickly into the engine to drive more efficient and superior performance. Onum delivers transformational advantages across three critical dimensions: ⚡ Speed: Delivers 5X more events per second than its nearest competitor and processes security and observability data in real-time versus legacy batch and store methods. ⚡ Cost: Smart filtering reduces data storage costs by 50% through intelligent optimization. ⚡ Superior Outcomes: Real-time pipeline detection starts before data enters the Falcon platform, delivering up to 70% faster incident response with 40% less ingestion overhead. The future of cybersecurity is here – built on real-time, high-fidelity data. 👉 Learn more:

CrowdStrike

12,767 Aufrufe • vor 10 Monaten

🚨#BREAKING: A Massive Rescue operation is underway as more than 75 to 100 people are stranded on a large broken sheet of ice 📌#UpperRedLake | #Minnesota Currently, a massive rescue operation is underway at Upper Red Lake in Minnesota, with numerous emergency personnel and other agencies on the scene working to rescue over 75 to 100 people stranded after a large ice chunk broke off from the main ice on Upper Red Lake. Emergency responders are evacuating from the southeast part of the ice in Red Lake, and several areas are affected. Officials say there are no reports of anyone in the water yet, as this situation is still developing.
0:12

🚨#BREAKING: A Massive Rescue operation is underway as more than 75 to 100 people are stranded on a large broken sheet of ice 📌#UpperRedLake | #Minnesota Currently, a massive rescue operation is underway at Upper Red Lake in Minnesota, with numerous emergency personnel and other agencies on the scene working to rescue over 75 to 100 people stranded after a large ice chunk broke off from the main ice on Upper Red Lake. Emergency responders are evacuating from the southeast part of the ice in Red Lake, and several areas are affected. Officials say there are no reports of anyone in the water yet, as this situation is still developing.

R A W S A L E R T S

2,417,146 Aufrufe • vor 2 Jahren

After nine long months of testing, the B18.1 Booster test article is finally being removed from the Super Crusher. Over the course of that campaign, it completed 16 cryogenic pressure tests — making it the most extensive structural qualification program we’ve seen to date, by a significant margin (followed closely by B18.3 and S39.1) To me, that level of rigor reflects more than just the technological leap between V2 and V3. It also signals a broader shift in testing philosophy in the Starship program. The changes made to Super Heavy introduce significant risk throughout every stage of pre-flight testing. Its almost equivalent to the extreme uncertainty during the Booster 7 testing campaign. With that in mind, it’s reassuring to see that those risks are being addressed with this level of discipline. 🎥: Avid Space
0:18

After nine long months of testing, the B18.1 Booster test article is finally being removed from the Super Crusher. Over the course of that campaign, it completed 16 cryogenic pressure tests — making it the most extensive structural qualification program we’ve seen to date, by a significant margin (followed closely by B18.3 and S39.1) To me, that level of rigor reflects more than just the technological leap between V2 and V3. It also signals a broader shift in testing philosophy in the Starship program. The changes made to Super Heavy introduce significant risk throughout every stage of pre-flight testing. Its almost equivalent to the extreme uncertainty during the Booster 7 testing campaign. With that in mind, it’s reassuring to see that those risks are being addressed with this level of discipline. 🎥: Avid Space

Zack Golden

36,170 Aufrufe • vor 5 Monaten

We facilitated over 36,000 passengers at Udhna Station on 19th October 2025, which is about 50% increase compared to the same day in 2024. All passengers boarded trains by 4 PM on 19th October and reached their home on Diwali day to celebrate Diwali with their families and loved ones. The coordinated efforts are taken to avoid unregulated crowd. A dedicated holding area was created and a large number of ticket counters are provided for the convenience of passengers. More than 1.2 lakh passengers travelled from Udhna in the last five days. The current status of platforms and the holding area is shown in following images
0:11

We facilitated over 36,000 passengers at Udhna Station on 19th October 2025, which is about 50% increase compared to the same day in 2024. All passengers boarded trains by 4 PM on 19th October and reached their home on Diwali day to celebrate Diwali with their families and loved ones. The coordinated efforts are taken to avoid unregulated crowd. A dedicated holding area was created and a large number of ticket counters are provided for the convenience of passengers. More than 1.2 lakh passengers travelled from Udhna in the last five days. The current status of platforms and the holding area is shown in following images

Ministry of Railways

22,759 Aufrufe • vor 8 Monaten

Ordinary Ukrainian cities: Kharkiv, Sumy, Dnipro. Ordinary life, and Russian strikes that should never have happened. Every day in Ukraine is marked by large-scale attacks with strike drones, mostly “Shaheds”—it was Iran that taught Russia how to produce such drones. Today, more than a dozen drones have already hit our Kharkiv, one of the largest cities in Ukraine. Ordinary residential buildings have been damaged, and so far 8 people have been wounded, including a 12-year-old girl. No country should have to go through this. And against this backdrop, speaking about easing pressure on Russia, lifting sanctions, and so on, is definitely inappropriate and unhelpful. Russia kills people every day and drags out this war. The U.S. proposal for an unconditional ceasefire has been on the table for half a month now. Pressure on Russia is needed to save lives and to make diplomacy work faster and more effectively. Without pressure on Russia, there will be no result.
0:13

Ordinary Ukrainian cities: Kharkiv, Sumy, Dnipro. Ordinary life, and Russian strikes that should never have happened. Every day in Ukraine is marked by large-scale attacks with strike drones, mostly “Shaheds”—it was Iran that taught Russia how to produce such drones. Today, more than a dozen drones have already hit our Kharkiv, one of the largest cities in Ukraine. Ordinary residential buildings have been damaged, and so far 8 people have been wounded, including a 12-year-old girl. No country should have to go through this. And against this backdrop, speaking about easing pressure on Russia, lifting sanctions, and so on, is definitely inappropriate and unhelpful. Russia kills people every day and drags out this war. The U.S. proposal for an unconditional ceasefire has been on the table for half a month now. Pressure on Russia is needed to save lives and to make diplomacy work faster and more effectively. Without pressure on Russia, there will be no result.

Volodymyr Zelenskyy / Володимир Зеленський

591,632 Aufrufe • vor 1 Jahr

CARBON DESKTOP BETA APP NOW LIVE 🔥 📥 The #CarbonBrowser Desktop Beta App is now available to download for PC and Mac OS! What to expect in Version 117.0.5938.0: - Faster Web Browsing than Brave, Chrome or Safari - Integrated AdBlock compatible on YouTube Ads - Supports your Favourite Browser Extensions/Plugins - Synchronised Homepage to Display Featured DApps - Enabled FFMPEG proprietary CODECS - Plus a lot more A new era of #web3 browser is about to begin with Carbon Browser and $CSIX LFG! #CSIX 🔥 #WebBrowser #Browser #DeFi #Crypto #CryptoNews #BNBChain #BNB
0:19

CARBON DESKTOP BETA APP NOW LIVE 🔥 📥 The #CarbonBrowser Desktop Beta App is now available to download for PC and Mac OS! What to expect in Version 117.0.5938.0: - Faster Web Browsing than Brave, Chrome or Safari - Integrated AdBlock compatible on YouTube Ads - Supports your Favourite Browser Extensions/Plugins - Synchronised Homepage to Display Featured DApps - Enabled FFMPEG proprietary CODECS - Plus a lot more A new era of #web3 browser is about to begin with Carbon Browser and $CSIX LFG! #CSIX 🔥 #WebBrowser #Browser #DeFi #Crypto #CryptoNews #BNBChain #BNB

Carbon Browser

262,709 Aufrufe • vor 2 Jahren

Open-Source Multi-Sensor Data Platform for Neural 3D Reconstruction and Physical AI [📍github] It handles cameras, LiDAR, radar, poses, calibrations & labels in one clean format. No more messy custom parsers. • Super efficient (non-redundant storage) • New .itar single-file format with lightning-fast random access • Streams straight from S3/GCS/Azure – perfect for huge datasets • Built-in converters for Waymo, ScanNet++ & more • Already powers NVIDIA NuRec, 3DGRUT & gsplat Saw this at Janick Martinez Esturo, thanks for sharing! Easy to try: • pip install nvidia-ncore • GitHub: • Docs & project page: NCore slashes data wrangling time, cuts storage waste, and makes large-scale neural 3D training faster and simpler than ever. A real standard for physical AI. ——- Weekly robotics and AI insights. Subscribe free:
0:17

Open-Source Multi-Sensor Data Platform for Neural 3D Reconstruction and Physical AI [📍github] It handles cameras, LiDAR, radar, poses, calibrations & labels in one clean format. No more messy custom parsers. • Super efficient (non-redundant storage) • New .itar single-file format with lightning-fast random access • Streams straight from S3/GCS/Azure – perfect for huge datasets • Built-in converters for Waymo, ScanNet++ & more • Already powers NVIDIA NuRec, 3DGRUT & gsplat Saw this at Janick Martinez Esturo, thanks for sharing! Easy to try: • pip install nvidia-ncore • GitHub: • Docs & project page: NCore slashes data wrangling time, cuts storage waste, and makes large-scale neural 3D training faster and simpler than ever. A real standard for physical AI. ——- Weekly robotics and AI insights. Subscribe free:

Ilir Aliu

32,717 Aufrufe • vor 2 Monaten

FastMap: Revisiting Dense and Scalable Structure from Motion "FASTMAP, a redesigned SfM framework, achieves fast, high-accuracy dense structure from motion. On large scenes with thousands of images, FASTMAP is up to one to two orders of magnitude faster than GLOMAP and COLMAP. ... Importantly, FASTMAP achieves efficiency improvements while keeping comparable performance. Extensive experiments on eight datasets demonstrate pose estimation accuracy and novel view synthesis quality close to GLOMAP and COLMAP. " Contributions: 1. For all the iterative nonlinear optimization problems involved, we design algorithms such that the computational complexity of each iteration is only linear in the number of image pairs, not keypoint pairs or 3D points. This includes replacing the traditional bundle adjustment [50] present in previous SfM frameworks with a novel re-weighting epipolar adjustment algorithm, which is much more efficient. 2. Throughout the entire framework, we formulate as many steps as possible as GPU-friendly dense tensor operations. This allows us to implement the entire method in PyTorch [39], which provides seamless GPU acceleration.
0:12

FastMap: Revisiting Dense and Scalable Structure from Motion "FASTMAP, a redesigned SfM framework, achieves fast, high-accuracy dense structure from motion. On large scenes with thousands of images, FASTMAP is up to one to two orders of magnitude faster than GLOMAP and COLMAP. ... Importantly, FASTMAP achieves efficiency improvements while keeping comparable performance. Extensive experiments on eight datasets demonstrate pose estimation accuracy and novel view synthesis quality close to GLOMAP and COLMAP. " Contributions: 1. For all the iterative nonlinear optimization problems involved, we design algorithms such that the computational complexity of each iteration is only linear in the number of image pairs, not keypoint pairs or 3D points. This includes replacing the traditional bundle adjustment [50] present in previous SfM frameworks with a novel re-weighting epipolar adjustment algorithm, which is much more efficient. 2. Throughout the entire framework, we formulate as many steps as possible as GPU-friendly dense tensor operations. This allows us to implement the entire method in PyTorch [39], which provides seamless GPU acceleration.

MrNeRF

15,233 Aufrufe • vor 1 Jahr