Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

🚨BREAKING NEWS🚨: Phala Shipped the First-Ever GPU TEE Benchmark! Benchmark Research Highlight: ✅ LLaMa 3, Microsoft Phi Models Tested ✅ Tests Performed on nVIDIA H100 ✅ With #TEE Mode On 📊 Key Results: • Up to 150 TPS for LLaMA-3-8B • Performance trade-off as low as 0.5%, up to... show more

Phala

141,208 subscribers

68,981 views • 2 years ago •via X (Twitter)

Science & Technology News & Politics Education #TEE

Anya Rossi• Live Now

Private livecam show

10 Comments

Phala Network2 years ago

👀📷 Watch the whole video:

Phala Network2 years ago

has just dropped an awesome report on Phala's groundbreaking GPU TEE Benchmark release! 📊🔥 Read here:

𓂀 Dr.Naveen 𓂀2 years ago

$PHA is the #AI #DEPIN darkhorse of this bullrun🔥

infaceAi2 years ago

@rendernetwork @ionet what's up, guys 🤔😉

Tesseract2 years ago

Great work!

Same same but different2 years ago

I don't understand this but this is intense

Prashant - ai/acc | bringing spheron revolution2 years ago

This one is quite a good team; keep up the good work & looking forward to giving it a spin as well.

NFTPerks 🇵🇹2 years ago

will definitely watch it

Web3 with Asianpett2 years ago

Now here's making a difference and providing #web3 solutions. Cheers @PhalaNetwork 👍

Trader26💲💲2 years ago

🚀🚀🚀🚀

Related Videos

Ondo Perps is live. Up to 20x leverage on equity perps is here, with tokenized stocks as collateral & up to $3 million in rewards. Ondo technology first revolutionized tokenized stocks & now powers perps markets for: ✅ Oil ✅ Intel ✅ Gold ✅ AMD ✅ Meta ✅ Tesla ✅ Silver ✅ Apple ✅ Circle ✅ Nvidia ✅ Oracle ✅ Netflix ✅ DRAM ✅ Micron ✅ US 100 ✅ US 500 ✅ SpaceX ✅ Palantir ✅ Amazon ✅ Strategy ✅ Alphabet ✅ Coinbase ✅ Microsoft ✅ Robinhood All available for permissionless trading 24/7/365. Ondo Perps is the first platform to allow tokenized stocks as collateral, available now for Pre-Alpha users. RWA perps can now trade on a platform built to deliver liquidity and capital efficiency on par with traditional derivatives exchanges. Traders can deploy capital in size, run sophisticated strategies, and seamlessly manage risk all on a single platform that eliminates the constraints of traditional derivatives venues. No expiry. No closing bell. Just institutional-grade execution on the world’s most-traded markets. Accessible now to billions globally outside the U.S., Panama, and other prohibited jurisdictions.

Ondo Perps is live. Up to 20x leverage on equity perps is here, with tokenized stocks as collateral & up to $3 million in rewards. Ondo technology first revolutionized tokenized stocks & now powers perps markets for: ✅ Oil ✅ Intel ✅ Gold ✅ AMD ✅ Meta ✅ Tesla ✅ Silver ✅ Apple ✅ Circle ✅ Nvidia ✅ Oracle ✅ Netflix ✅ DRAM ✅ Micron ✅ US 100 ✅ US 500 ✅ SpaceX ✅ Palantir ✅ Amazon ✅ Strategy ✅ Alphabet ✅ Coinbase ✅ Microsoft ✅ Robinhood All available for permissionless trading 24/7/365. Ondo Perps is the first platform to allow tokenized stocks as collateral, available now for Pre-Alpha users. RWA perps can now trade on a platform built to deliver liquidity and capital efficiency on par with traditional derivatives exchanges. Traders can deploy capital in size, run sophisticated strategies, and seamlessly manage risk all on a single platform that eliminates the constraints of traditional derivatives venues. No expiry. No closing bell. Just institutional-grade execution on the world’s most-traded markets. Accessible now to billions globally outside the U.S., Panama, and other prohibited jurisdictions.

Ondo Perps

727,002 views • 1 month ago

🚨NEW INSTANT PRO ACCOUNT🚨 GFT introduces the first instant plan with NO DAILY DRAWDOWN! Here’s what’s on the table: ✅ No Daily DD ✅ 1:50 Leverage ✅ Up to 100% Profit Split ✅ Bi-weekly or on Demand Payouts Enjoy the launch promo: 🔥40% OFF SITEWIDE 🎟️ Code: 40SPECIAL Available for a limited time only! • 2,5K Instant Pro Plan for just $42 • 100K Instant Pro Plan for just $489 Skip the Challenge and start trading right away Trade like the greatest 🏆

🚨NEW INSTANT PRO ACCOUNT🚨 GFT introduces the first instant plan with NO DAILY DRAWDOWN! Here’s what’s on the table: ✅ No Daily DD ✅ 1:50 Leverage ✅ Up to 100% Profit Split ✅ Bi-weekly or on Demand Payouts Enjoy the launch promo: 🔥40% OFF SITEWIDE 🎟️ Code: 40SPECIAL Available for a limited time only! • 2,5K Instant Pro Plan for just $42 • 100K Instant Pro Plan for just $489 Skip the Challenge and start trading right away Trade like the greatest 🏆

Goat Funded Trader

41,973 views • 11 months ago

Check out mistral.rs, our #Rust-based open source inference engine allowing for fast #LLM serving for a variety of architectures including X-LoRA mixture-of-expert (MoE) models, Llama-3, Mistral/Mixtral, Gemma & many others. Built on the Hugging Face #Candle framework for #Rust w/ custom CUDA kernels in the backend (as well as support for Metal, Apple Accelerate, and Intel MKL for CPU use), you can easily create a REST API OpenAI compatible server or run via Python bindings. Key features include: ✅Prefix caching, continuous batching ✅Flash Attention V2 ✅Device offloading ✅GGUF or Hugging Face models ✅2, 3, 4, 5, 6 and 8 bit quantization ✅X-LoRA MoE non-granular scalings for fast inference ✅Grammar support ✅Continuous batching ✅LoRA support with weight merging ✅LlamaIndex 🦙 integration ...and much more. Incorporation into our GraphReasoning multi-agent modeling framework & LlamaIndex 🦙 allows you to combine in-context learning with adversarial agentic strategies, to dive deep into complex scientific analyses, such as to predict material behaviors, generate hypotheses, analyze papers and data, develop new research concepts, and much more. Check out mistral.rs: Join our Discord here: Rust Trending Rust Language

Check out mistral.rs, our #Rust-based open source inference engine allowing for fast #LLM serving for a variety of architectures including X-LoRA mixture-of-expert (MoE) models, Llama-3, Mistral/Mixtral, Gemma & many others. Built on the Hugging Face #Candle framework for #Rust w/ custom CUDA kernels in the backend (as well as support for Metal, Apple Accelerate, and Intel MKL for CPU use), you can easily create a REST API OpenAI compatible server or run via Python bindings. Key features include: ✅Prefix caching, continuous batching ✅Flash Attention V2 ✅Device offloading ✅GGUF or Hugging Face models ✅2, 3, 4, 5, 6 and 8 bit quantization ✅X-LoRA MoE non-granular scalings for fast inference ✅Grammar support ✅Continuous batching ✅LoRA support with weight merging ✅LlamaIndex 🦙 integration ...and much more. Incorporation into our GraphReasoning multi-agent modeling framework & LlamaIndex 🦙 allows you to combine in-context learning with adversarial agentic strategies, to dive deep into complex scientific analyses, such as to predict material behaviors, generate hypotheses, analyze papers and data, develop new research concepts, and much more. Check out mistral.rs: Join our Discord here: Rust Trending Rust Language

Markus J. Buehler

73,581 views • 2 years ago

🚨 #BREAKING: President Trump announces a MAJOR trade deal with the United Kingdom This ended up being a HELL of a deal for us! ✅ $6 BILLION in new tariff revenue ✅ US tariffs on imports from UK raised to 10% ✅ UK tariffs on imports from US lowered to 1.8% ✅ $5 BILLION in new exports from the U.S. “The deal includes billions of dollars of increased market access for American exports, especially in agriculture, dramatically increasing access for American beef, ethanol, and virtually all of the products produced by our great farmers.”

🚨 #BREAKING: President Trump announces a MAJOR trade deal with the United Kingdom This ended up being a HELL of a deal for us! ✅ $6 BILLION in new tariff revenue ✅ US tariffs on imports from UK raised to 10% ✅ UK tariffs on imports from US lowered to 1.8% ✅ $5 BILLION in new exports from the U.S. “The deal includes billions of dollars of increased market access for American exports, especially in agriculture, dramatically increasing access for American beef, ethanol, and virtually all of the products produced by our great farmers.”

Nick Sortor

2,495,662 views • 1 year ago

$1M IN $NODO Airdrop ALERT– Earn XP & GEMs Before TGE! 🚨 Early adopters win BIG! Stack XP & GEMs now to claim $NODO at TGE. Here’s how: NODO XP = Free $NODO for community builders 💡 Ways to earn XP 👇 ✅ Connect wallet on NODO - 10,000 XP ✅ Follow NODO on Twitter & Telegram – 10,000 XP each ✅ Deposit & Earn – Get 1000 XP per $1 deposited ✅ Predict & Earn – Trade in the Prediction Market & get 5,000 XP per $1 ✅ Refer & Earn – Get 30% of your friend’s XP & 10% from their referrals (up to 3 levels!) NODO GEMs = Premium perks for traders 💎 Ways to earn GEMs 👇 ✅ Sign Up Bonus – Get 1,500 GEMs & 1,200 XP just for joining ✅ Trade & Earn – Exclusive GEM rewards with every trade ✅ Leaderboard Bonus – Rank Top 10 weekly by trading on NODO & win extra GEMs ⏳ The earlier you start, the bigger your claim! Don’t get left behind—secure your share now! 🚀 👉 Join the airdrop:

$1M IN $NODO Airdrop ALERT– Earn XP & GEMs Before TGE! 🚨 Early adopters win BIG! Stack XP & GEMs now to claim $NODO at TGE. Here’s how: NODO XP = Free $NODO for community builders 💡 Ways to earn XP 👇 ✅ Connect wallet on NODO - 10,000 XP ✅ Follow NODO on Twitter & Telegram – 10,000 XP each ✅ Deposit & Earn – Get 1000 XP per $1 deposited ✅ Predict & Earn – Trade in the Prediction Market & get 5,000 XP per $1 ✅ Refer & Earn – Get 30% of your friend’s XP & 10% from their referrals (up to 3 levels!) NODO GEMs = Premium perks for traders 💎 Ways to earn GEMs 👇 ✅ Sign Up Bonus – Get 1,500 GEMs & 1,200 XP just for joining ✅ Trade & Earn – Exclusive GEM rewards with every trade ✅ Leaderboard Bonus – Rank Top 10 weekly by trading on NODO & win extra GEMs ⏳ The earlier you start, the bigger your claim! Don’t get left behind—secure your share now! 🚀 👉 Join the airdrop:

NODO

26,152 views • 1 year ago

NVIDIA AI Releases Nemotron 3 Embed: An Open Embedding Collection Whose 8B Checkpoint Ranks #1 on RTEB Most RAG stacks treat the embedding model as a commodity — pick one, index, move on. Nemotron 3 Embed is NVIDIA's argument that the retrieval layer is where agent cost actually gets set. They released three open checkpoints — Nemotron-3-Embed-8B-BF16, 1B-BF16, and 1B-NVFP4 — built on Ministral bases, trained with bidirectional attention masking, pooled by averaging token-level representations, all taking 32,768-token inputs under OpenMDW-1.1. Here's what's actually interesting: → The 8B ranks #1 overall on RTEB: 78.46 avg NDCG@10, alongside 75.45 on MMTEB Retrieval and 60.60 on ViDoRe-V3 text → The 1B wasn't trained small. It was pruned from a 3B parent with ModelOpt mcore_minitron NAS, then distilled from the 8B teacher on COS + MSE loss — twice → That pipeline lands the 1.14B checkpoint at 72.38 RTEB, up 10.4 points on the prior-generation llama-nemotron-embed-vl-1b-v2 → NVFP4 costs 0.38 RTEB points (72.00 vs 72.38, ~99.5% retention) and buys up to 2x BF16 throughput on Blackwell Full Analysis: Model weight: Technical details: NVIDIA AI NVIDIA

NVIDIA AI Releases Nemotron 3 Embed: An Open Embedding Collection Whose 8B Checkpoint Ranks #1 on RTEB Most RAG stacks treat the embedding model as a commodity — pick one, index, move on. Nemotron 3 Embed is NVIDIA's argument that the retrieval layer is where agent cost actually gets set. They released three open checkpoints — Nemotron-3-Embed-8B-BF16, 1B-BF16, and 1B-NVFP4 — built on Ministral bases, trained with bidirectional attention masking, pooled by averaging token-level representations, all taking 32,768-token inputs under OpenMDW-1.1. Here's what's actually interesting: → The 8B ranks #1 overall on RTEB: 78.46 avg NDCG@10, alongside 75.45 on MMTEB Retrieval and 60.60 on ViDoRe-V3 text → The 1B wasn't trained small. It was pruned from a 3B parent with ModelOpt mcore_minitron NAS, then distilled from the 8B teacher on COS + MSE loss — twice → That pipeline lands the 1.14B checkpoint at 72.38 RTEB, up 10.4 points on the prior-generation llama-nemotron-embed-vl-1b-v2 → NVFP4 costs 0.38 RTEB points (72.00 vs 72.38, ~99.5% retention) and buys up to 2x BF16 throughput on Blackwell Full Analysis: Model weight: Technical details: NVIDIA AI NVIDIA

Marktechpost AI

32,352 views • 1 month ago

🚨 MAJOR Internet Money Wallet is live on Android, iOS and Chrome. ✅ By consistent community request we have updated the overall look (User Interface/UI). 📝This is Phase 3 of our UI overhaul. More to come. 🤔Watch the video below for a more in-depth description. Some of the updates include: ✅New portfolio change feature on the home and wallet screens ✅Updated icons app wide ✅Updated color continuity and philosophy app wide ✅Updated fonts app wide ✅Reconstructed the Home and wallet screens ✅Reconstructed the navigation (nav) bars on both mobile and chrome ✅Added filtering options for Transactions, Home, Wallet, Networks and NFTs ✅Fixed several bugs ✅Implemented several speed optimizations 🔔Hit the bell icon on our page to stay up to date as we announce future releases. 🧠 Make sure your Chrome is on version: 2.3.0 and your Android and iOS devices are on version 2.2.10 to see these awesome updates. 🙏 Thank you to our incredible community who continues to use Internet Money and consistently provides the feedback necessary to grow and improve. As well as onboards their friends and family!

🚨 MAJOR Internet Money Wallet is live on Android, iOS and Chrome. ✅ By consistent community request we have updated the overall look (User Interface/UI). 📝This is Phase 3 of our UI overhaul. More to come. 🤔Watch the video below for a more in-depth description. Some of the updates include: ✅New portfolio change feature on the home and wallet screens ✅Updated icons app wide ✅Updated color continuity and philosophy app wide ✅Updated fonts app wide ✅Reconstructed the Home and wallet screens ✅Reconstructed the navigation (nav) bars on both mobile and chrome ✅Added filtering options for Transactions, Home, Wallet, Networks and NFTs ✅Fixed several bugs ✅Implemented several speed optimizations 🔔Hit the bell icon on our page to stay up to date as we announce future releases. 🧠 Make sure your Chrome is on version: 2.3.0 and your Android and iOS devices are on version 2.2.10 to see these awesome updates. 🙏 Thank you to our incredible community who continues to use Internet Money and consistently provides the feedback necessary to grow and improve. As well as onboards their friends and family!

Internet Money | Crypto Wallet

21,181 views • 6 months ago

We’re launching Optima. Now anyone can create a custom benchmark for their use case, leveraging Artificial Analysis’ leading research and platform Building and running benchmarks is difficult. We have distilled Artificial Analysis’ research and experience developing benchmarks into Optima, a new platform for benchmarking models on your own workloads and comparing performance, speed and cost efficiency. Optima allows you to find the best model for your task, or an equally performant alternative to your current setup at 10x lower cost or time per task. We’ve integrated Artificial Analysis' research and experience in benchmarks across the Optima workflow: ➤ Build benchmarks based on your own data and use cases: There are three ways to build a benchmark with Optima. Upload an existing evaluation dataset from your own files or Hugging Face, or import agent traces from platforms including Arize AI, Braintrust and langfuse.com. Install the Optima skill to build a benchmark using context from your coding environment and previous sessions. Or simply describe your use case and provide example inputs and outputs, and Optima will build the benchmark for you ➤ Run across the latest models: Run the same benchmark across leading models in a single click, and keep your leaderboard up to date as soon as new models are released ➤ Bring Artificial Analysis grading to your own benchmark: Evaluate responses against objective rubric criteria or using the same pairwise judging approach used for Artificial Analysis benchmarks including GDPval-AA and AA-Briefcase. For pairwise judging, select your preferred responses from a sample and Optima uses those preferences to rank models across your test set ➤ Compare performance, cost and time efficiency: Optima measures more than model performance. Cost per Task and Time per Task are tracked alongside benchmark scores, with category-level results and support for custom metrics, allowing you to compare the tradeoffs between models for your specific use case Ahead of launch, here are examples questions our beta testers answered with Optima: ➤ Which model can save me 10x the cost without a meaningful decrease in quality for my finance & accounting agent? ➤ Which model best matches the writing style of lawyers for my legal agent? ➤ Which model can best identify different elements in my custom image dataset? Optima is available today. Build your own benchmark at

We’re launching Optima. Now anyone can create a custom benchmark for their use case, leveraging Artificial Analysis’ leading research and platform Building and running benchmarks is difficult. We have distilled Artificial Analysis’ research and experience developing benchmarks into Optima, a new platform for benchmarking models on your own workloads and comparing performance, speed and cost efficiency. Optima allows you to find the best model for your task, or an equally performant alternative to your current setup at 10x lower cost or time per task. We’ve integrated Artificial Analysis' research and experience in benchmarks across the Optima workflow: ➤ Build benchmarks based on your own data and use cases: There are three ways to build a benchmark with Optima. Upload an existing evaluation dataset from your own files or Hugging Face, or import agent traces from platforms including Arize AI, Braintrust and langfuse.com. Install the Optima skill to build a benchmark using context from your coding environment and previous sessions. Or simply describe your use case and provide example inputs and outputs, and Optima will build the benchmark for you ➤ Run across the latest models: Run the same benchmark across leading models in a single click, and keep your leaderboard up to date as soon as new models are released ➤ Bring Artificial Analysis grading to your own benchmark: Evaluate responses against objective rubric criteria or using the same pairwise judging approach used for Artificial Analysis benchmarks including GDPval-AA and AA-Briefcase. For pairwise judging, select your preferred responses from a sample and Optima uses those preferences to rank models across your test set ➤ Compare performance, cost and time efficiency: Optima measures more than model performance. Cost per Task and Time per Task are tracked alongside benchmark scores, with category-level results and support for custom metrics, allowing you to compare the tradeoffs between models for your specific use case Ahead of launch, here are examples questions our beta testers answered with Optima: ➤ Which model can save me 10x the cost without a meaningful decrease in quality for my finance & accounting agent? ➤ Which model best matches the writing style of lawyers for my legal agent? ➤ Which model can best identify different elements in my custom image dataset? Optima is available today. Build your own benchmark at

Artificial Analysis

130,269 views • 22 days ago

A team tested Pi0, Pi0 Fast, Gr00t, and ACT on real robot arms in manufacturing tasks. (🔖 Bookmark this for later!) The task was precise: place thin rectangular frames from a messy stack into a holder. The team fine-tuned each model on 100 real trajectories and compared training time, inference speed, motion quality, and success rates. ⬇️ Here’s a breakdown of what they found Pi0 (Original) ✅ Strongest overall performance in precise pick-and-place ✅ High success rate even in edge cases ✅ Longest training time (~11 hours, ~$30 per run) ✅ Inference time of 80 ms causes short pauses between actions Despite delays, it handles complex scenarios well… solid for high-precision tasks, but slow to train. Gr00t ✅ Trains fast (~2 hours, ~$5 per run) ✅ Performs almost as well as Pi0 on large-object tasks ✅ Struggles with fine precision; random movement in some trials ✅ More training didn’t fix jitter or random offsets Best suited for tasks where exact precision isn’t critical. Not ready for manufacturing-grade accuracy without more tuning. Pi0 Fast ✅ Promised faster training, but results were underwhelming ✅ Training at 6 hours still showed low success rates ✅ Inference was slower than expected ✅ Not reliable for generalizing even slightly new tasks Currently too unstable for real-world deployment. Doesn’t live up to the “Fast” name yet. ACT (Baseline) ✅ 200MB model—lightweight, but limited ✅ Struggles with stacked objects or ambiguous scenes ✅ Success rates around 70% in best-case setups ✅ Can’t match newer models on precision or generalization Still a solid baseline, but clearly a generation behind in robustness. 🚨 Extra Notes All newer models share a common issue: •Inference takes longer than a frame (80 ms vs 33 ms), so robots “pause” between chunks. •This results in jittery movements, but not a dealbreaker unless tasks are time-sensitive. Language-conditioned tasks also fell short: after training on two labeled tasks, the model couldn’t generalize to a third unseen combination using only text prompts. ✅ The good news? These models adapt well to new robot arms with quick fine-tuning. ❌ The bad news? There’s still no plug-and-play solution for improving performance after deployment. Reinforcement learning or DAgger-style data collection during real-world operation may be the next big step, something many teams in robotics are actively working on.

A team tested Pi0, Pi0 Fast, Gr00t, and ACT on real robot arms in manufacturing tasks. (🔖 Bookmark this for later!) The task was precise: place thin rectangular frames from a messy stack into a holder. The team fine-tuned each model on 100 real trajectories and compared training time, inference speed, motion quality, and success rates. ⬇️ Here’s a breakdown of what they found Pi0 (Original) ✅ Strongest overall performance in precise pick-and-place ✅ High success rate even in edge cases ✅ Longest training time (~11 hours, ~$30 per run) ✅ Inference time of 80 ms causes short pauses between actions Despite delays, it handles complex scenarios well… solid for high-precision tasks, but slow to train. Gr00t ✅ Trains fast (~2 hours, ~$5 per run) ✅ Performs almost as well as Pi0 on large-object tasks ✅ Struggles with fine precision; random movement in some trials ✅ More training didn’t fix jitter or random offsets Best suited for tasks where exact precision isn’t critical. Not ready for manufacturing-grade accuracy without more tuning. Pi0 Fast ✅ Promised faster training, but results were underwhelming ✅ Training at 6 hours still showed low success rates ✅ Inference was slower than expected ✅ Not reliable for generalizing even slightly new tasks Currently too unstable for real-world deployment. Doesn’t live up to the “Fast” name yet. ACT (Baseline) ✅ 200MB model—lightweight, but limited ✅ Struggles with stacked objects or ambiguous scenes ✅ Success rates around 70% in best-case setups ✅ Can’t match newer models on precision or generalization Still a solid baseline, but clearly a generation behind in robustness. 🚨 Extra Notes All newer models share a common issue: •Inference takes longer than a frame (80 ms vs 33 ms), so robots “pause” between chunks. •This results in jittery movements, but not a dealbreaker unless tasks are time-sensitive. Language-conditioned tasks also fell short: after training on two labeled tasks, the model couldn’t generalize to a third unseen combination using only text prompts. ✅ The good news? These models adapt well to new robot arms with quick fine-tuning. ❌ The bad news? There’s still no plug-and-play solution for improving performance after deployment. Reinforcement learning or DAgger-style data collection during real-world operation may be the next big step, something many teams in robotics are actively working on.

Ilir Aliu

21,844 views • 1 year ago

🚀 OpenLLM 0.6 is out providing fast inference speed! Check out our video comparing #OpenLLM and #Ollama handling concurrent requests on the Llama 3 8B model. Ollama is ideal for local LLM deployment, but it is not designed for high concurrency scenarios essential for deployments in the cloud, powering applications with a large number of users. OpenLLM enables you to run any open-source model as OpenAI-compatible endpoints with just a single command. It supports a broad array of models like Llama 3, Mistral, and Qwen 2, including those fine-tuned with proprietary datasets and quantized versions. Ready to speed up your AI deployments? See the OpenLLM repo #BentoML #OpenSource

🚀 OpenLLM 0.6 is out providing fast inference speed! Check out our video comparing #OpenLLM and #Ollama handling concurrent requests on the Llama 3 8B model. Ollama is ideal for local LLM deployment, but it is not designed for high concurrency scenarios essential for deployments in the cloud, powering applications with a large number of users. OpenLLM enables you to run any open-source model as OpenAI-compatible endpoints with just a single command. It supports a broad array of models like Llama 3, Mistral, and Qwen 2, including those fine-tuned with proprietary datasets and quantized versions. Ready to speed up your AI deployments? See the OpenLLM repo #BentoML #OpenSource

BentoML

18,798 views • 2 years ago

This one really means a lot to me. Huge shoutout to everyone who's been riding with me since 2019 when I created the first ever realistics map through Fortnite Developers To thank you all, I've pushed myself harder than ever and completely updated 3 of my most popular and featured realistics maps that have over 2 billion plays. Some new key features: ✅ Speed Mode Enabled by Default ✅ 27 Brand New POIs ✅ Ranks & Top 250 Leaderboard Updated Daily ✅ Latest Loot Pool at ALL Times 🏆1V1: 7950-6306-4857 🏆2V2: 6570-5231-1418 🏆3V3: 6369-4500-6726 Also: There's a $1,000 contest up for the first 50 players to hit diamond rank on these maps. Refer to the comments for the details! 👇

Finest Creative

231,600 views • 2 years ago

Web scraping just leveled up Scrapling bypasses Cloudflare blocks, is 774 times faster than BeautifulSoup, and doesn't require proxy setup 52.2k stars on GitHub It's not just another scraper It's an adaptive framework that learns the structure of each website and automatically adjusts when it changes No manual maintenance. No getting blocked. ✅ Bypasses Cloudflare and the most aggressive anti-bots ✅ 774x faster than BeautifulSoup in real benchmarks ✅ No need for proxies or special setup ✅ Automatically adapts when the website structure changes ✅ Compatible with AI agents as an MCP server ✅ Support for JavaScript, iframes, and dynamic content ✅ Stealth mode for sites with advanced detection ✅ 46 releases. Updated last week. ✅ BSD-3 license What used to take you days to build and maintain now takes minutes 52.2k stars. 5k forks. BSD-3. repo here 👇

Web scraping just leveled up Scrapling bypasses Cloudflare blocks, is 774 times faster than BeautifulSoup, and doesn't require proxy setup 52.2k stars on GitHub It's not just another scraper It's an adaptive framework that learns the structure of each website and automatically adjusts when it changes No manual maintenance. No getting blocked. ✅ Bypasses Cloudflare and the most aggressive anti-bots ✅ 774x faster than BeautifulSoup in real benchmarks ✅ No need for proxies or special setup ✅ Automatically adapts when the website structure changes ✅ Compatible with AI agents as an MCP server ✅ Support for JavaScript, iframes, and dynamic content ✅ Stealth mode for sites with advanced detection ✅ 46 releases. Updated last week. ✅ BSD-3 license What used to take you days to build and maintain now takes minutes 52.2k stars. 5k forks. BSD-3. repo here 👇

Indu Tripathi

20,815 views • 3 months ago

Good news and bad news 😖🤧: Hello guys, in the spirit of transparency, I have good news & bad news on the attempt on full power tests on the mark3 prototype. The bad news is, the driving motor shaft developed a critical fault & therefore we could not reach high speeds for the engine to self sustain to idle as before🤦🏻‍♂️, the shaft eventually broke off as shown in the video❌. The good news is, our afterburner sequence works✅✅ even below self sustaining speeds, I decided to activate it at this low power level & it worked 🔥, although at full power it would be more violent & streamlined in nature, the sun's glare also made the flames less visible. We will fix the motor damage & try full power testing again next week 🙌🏻. I am not giving up, we are not giving up ⚠️, we are almost there, full power or nothing 🚀 Thank you for the kindness and support, I appreciate you all, we will fix the anomaly & test again. I feel very sad but goodnight 🙏🏻 #Zedora

Good news and bad news 😖🤧: Hello guys, in the spirit of transparency, I have good news & bad news on the attempt on full power tests on the mark3 prototype. The bad news is, the driving motor shaft developed a critical fault & therefore we could not reach high speeds for the engine to self sustain to idle as before🤦🏻‍♂️, the shaft eventually broke off as shown in the video❌. The good news is, our afterburner sequence works✅✅ even below self sustaining speeds, I decided to activate it at this low power level & it worked 🔥, although at full power it would be more violent & streamlined in nature, the sun's glare also made the flames less visible. We will fix the motor damage & try full power testing again next week 🙌🏻. I am not giving up, we are not giving up ⚠️, we are almost there, full power or nothing 🚀 Thank you for the kindness and support, I appreciate you all, we will fix the anomaly & test again. I feel very sad but goodnight 🙏🏻 #Zedora

Ziko Abara

111,756 views • 1 year ago

In just one week, Binh and I trained a full-body Unitree G1. Here's a recap: 1. Secured a Unitree G1 humanoid through a LinkedIn post 2. Deployed TWIST2 full-body teleoperation pipelines 3. Adapted TWIST2 for Zed stereo camera & collected full-body teleoperation samples (carried by Binh ) 4. Adapted & fine-tuned NVIDIA Gr00T N1.5 VLA on the TWIST2 public datasets, which I fine-tuned on an 8xNVIDIA H100 Cluster. We picked Gr00T N1.5 as it was trained with Unitree G1 embodiment data. 5. Adapted the TWIST2 codebase to stream in the actions from Gr00T via ZMQ using a co-located NVIDIA H100 for ~200ms inference latency 6. Tested the model in sim, then deployed to the real-world Unitree G1. We streamed a training sample observation to the VLA (as we didn't want to break robot in case real observations were OOD) We were the first team in the world to deploy the full TWIST2 data collection pipeline to the unitree g1 :) Much more work ahead though, which I'll work on as a side-project over the next months: 1. Exploring the various types of 'world models': video backbones, dynamics models, v-jepa-2 models. I believe these will generalize better & train much more data-efficiently than VLM backbones 2. Speeding up inference - I believe low-latency robotics inference will be a big challenge. There are many works in video diffusion which I'd like to test (e.g. SageAttention, SparseAttention, Drifting Models). Perhaps also writing custom CUDA kernels. 3. Economics of inference scaling :) What will be the compute demands as we scale inference up to millions of humanoids? Will it run on edge or on distributed 'co-located' inference clusters? These are questions I'd like to answer. Adapted TWIST2 codebase: Adapted Gr00T-N1.5 codebase: The ETH Robotics Club are doing a cool GTC Golden ticket competition with NVIDIA , so this is my submission :) The DGX Spark compute will get me a long way with initial prototyping & especially working on inference optimization for next-gen Blackwell GPUs #NVIDIAGTC #GOLDENTICKET #ETHRC

In just one week, Binh and I trained a full-body Unitree G1. Here's a recap: 1. Secured a Unitree G1 humanoid through a LinkedIn post 2. Deployed TWIST2 full-body teleoperation pipelines 3. Adapted TWIST2 for Zed stereo camera & collected full-body teleoperation samples (carried by Binh ) 4. Adapted & fine-tuned NVIDIA Gr00T N1.5 VLA on the TWIST2 public datasets, which I fine-tuned on an 8xNVIDIA H100 Cluster. We picked Gr00T N1.5 as it was trained with Unitree G1 embodiment data. 5. Adapted the TWIST2 codebase to stream in the actions from Gr00T via ZMQ using a co-located NVIDIA H100 for ~200ms inference latency 6. Tested the model in sim, then deployed to the real-world Unitree G1. We streamed a training sample observation to the VLA (as we didn't want to break robot in case real observations were OOD) We were the first team in the world to deploy the full TWIST2 data collection pipeline to the unitree g1 :) Much more work ahead though, which I'll work on as a side-project over the next months: 1. Exploring the various types of 'world models': video backbones, dynamics models, v-jepa-2 models. I believe these will generalize better & train much more data-efficiently than VLM backbones 2. Speeding up inference - I believe low-latency robotics inference will be a big challenge. There are many works in video diffusion which I'd like to test (e.g. SageAttention, SparseAttention, Drifting Models). Perhaps also writing custom CUDA kernels. 3. Economics of inference scaling :) What will be the compute demands as we scale inference up to millions of humanoids? Will it run on edge or on distributed 'co-located' inference clusters? These are questions I'd like to answer. Adapted TWIST2 codebase: Adapted Gr00T-N1.5 codebase: The ETH Robotics Club are doing a cool GTC Golden ticket competition with NVIDIA , so this is my submission :) The DGX Spark compute will get me a long way with initial prototyping & especially working on inference optimization for next-gen Blackwell GPUs #NVIDIAGTC #GOLDENTICKET #ETHRC

Arnie Ramesh

23,236 views • 6 months ago

🚀 Join the NFT Dynamic Ambassador & Referral Program – Unlock Exclusive Rewards! 💼 3 Tiers of Opportunity: 👑 Evangelist (Special Ambassador) – Elite Status, Exclusive Rewards! ✅ Limited to 22 early buyers on 3rd round ✅ Earn 10% commission on direct referrals ✅ Airdrop bonuses for 100+ referrals ⚠️ Maintain status by referring 2 users every 3 days 🌟 Ambassador – Grow & Earn! ✅ Upgrade from Investor after referring 10 users ✅ 5% discount on NFT purchases via referral ✅ 5% commission on direct referrals ⚠️ Maintain status by referring 2 users every 6 days 💰 Investor – Your First Step! ✅ Earn commissions & level up to Ambassador! 🔥 NFT Issuance & Trading – Own, Stake, Trade! 🔹3rd Phase: 22 NFTs for Evangelist status, 12h22 GMT today. 🔹 Smart Contract Minting: Auto-generated NFTs with referral tracking 🔹 Stake Your NFT to unlock VIP benefits & high interest rates 🔹 Seamless Trading: Fully transferable with all perks intact! 💎 Start small, build BIG – The earlier you start, the greater the rewards! 🔥 The 3rd Evangelist NFT Dynamic Round’ll be OPEN SOON - Stay tuned!

🚀 Join the NFT Dynamic Ambassador & Referral Program – Unlock Exclusive Rewards! 💼 3 Tiers of Opportunity: 👑 Evangelist (Special Ambassador) – Elite Status, Exclusive Rewards! ✅ Limited to 22 early buyers on 3rd round ✅ Earn 10% commission on direct referrals ✅ Airdrop bonuses for 100+ referrals ⚠️ Maintain status by referring 2 users every 3 days 🌟 Ambassador – Grow & Earn! ✅ Upgrade from Investor after referring 10 users ✅ 5% discount on NFT purchases via referral ✅ 5% commission on direct referrals ⚠️ Maintain status by referring 2 users every 6 days 💰 Investor – Your First Step! ✅ Earn commissions & level up to Ambassador! 🔥 NFT Issuance & Trading – Own, Stake, Trade! 🔹3rd Phase: 22 NFTs for Evangelist status, 12h22 GMT today. 🔹 Smart Contract Minting: Auto-generated NFTs with referral tracking 🔹 Stake Your NFT to unlock VIP benefits & high interest rates 🔹 Seamless Trading: Fully transferable with all perks intact! 💎 Start small, build BIG – The earlier you start, the greater the rewards! 🔥 The 3rd Evangelist NFT Dynamic Round’ll be OPEN SOON - Stay tuned!

AIIA DeFi

14,106 views • 1 year ago

Heavy RDLs🚂 Working with 500x9 here These are another exercise you can never really be too strong on. I keep them in rotation as a main accessory for carryover to both squat AND deadlift. A few coaching cues👇 ✅Hinge back at the hips FIRST. This loads the glutes- Many lifters mess this up by hinging at the waist instead, turning it into purely a low back exercise. ✅Keep chin tucked. This helps create a neutral spine which again, is key for loading the glutes ✅Allow a slight knee bend as you bring the bar down to allow the hips to go back further. ✅Descend until you feel the load fully in your glutes, then push the hips forward and stand back up. Bookmark this and share it out if you found this helpful!

Heavy RDLs🚂 Working with 500x9 here These are another exercise you can never really be too strong on. I keep them in rotation as a main accessory for carryover to both squat AND deadlift. A few coaching cues👇 ✅Hinge back at the hips FIRST. This loads the glutes- Many lifters mess this up by hinging at the waist instead, turning it into purely a low back exercise. ✅Keep chin tucked. This helps create a neutral spine which again, is key for loading the glutes ✅Allow a slight knee bend as you bring the bar down to allow the hips to go back further. ✅Descend until you feel the load fully in your glutes, then push the hips forward and stand back up. Bookmark this and share it out if you found this helpful!

Jeromy Bryk | Strength+Physique Coach💯

15,882 views • 8 months ago

Scaling vision-language-action (VLA) models to high-DoF dexterous hands has long been a "holy grail" challenge due to the high-dimensional action space and data scarcity. As a wrap up of the year 2025, we are releasing GR-Dexter, a holistic hardware-model-data framework for generalist manipulation on a bimanual dexterous-hand robot. This is the first VLA system to achieve: ✅ High-DoF Control: Managing a 56-DoF bimanual system (21-DoF per hand). ✅ Long-Horizon Tasks with tool use: Vacuuming, bread serving with tongs, and table decluttering. ✅ Open-World Generalization: Robust performance with unseen objects and abstract instructions. Project page: ArXiv:

Scaling vision-language-action (VLA) models to high-DoF dexterous hands has long been a "holy grail" challenge due to the high-dimensional action space and data scarcity. As a wrap up of the year 2025, we are releasing GR-Dexter, a holistic hardware-model-data framework for generalist manipulation on a bimanual dexterous-hand robot. This is the first VLA system to achieve: ✅ High-DoF Control: Managing a 56-DoF bimanual system (21-DoF per hand). ✅ Long-Horizon Tasks with tool use: Vacuuming, bread serving with tongs, and table decluttering. ✅ Open-World Generalization: Robust performance with unseen objects and abstract instructions. Project page: ArXiv:

Xiao Ma

94,056 views • 8 months ago