Distributed training on M4 Mac Mini cluster We implemented... Google DeepMind DiLoCo on Apple Silicon to train large models with 100-1000x less bandwidth compared to DDP baseline. AI is entering a new era where a distributed network of consumer devices can train large models.show more

Running DeepSeek-V3 on M4 Mac Mini AI Cluster 671B... MoE model distributed across 8 M4 Pro 64GB Mac Minis. Apple Silicon with unified memory is a great fit for MoE.show more

A teaser of DeepCoreCloud! With 100+ GPUs in a... high-performance cluster, we’re bringing 24/7 AI computing to the world. Train models, deploy AI, and scale without limits! DeepCoreCloud is built for innovators. The future of AI infrastructure is almost here. Are you ready? #DeepCoreCloud #ItronixAI #GPUClustershow more

Canter 2B is a photography-oriented text-to-image model being trained... on just a single GPU by data-archetype 🐎 are we entering the indie ai models era? still training, the v0002 checkpoint is available, more to come ▶️ play on spacesshow more

As a supporter of the open weights ecosystem, we're... proud to be a post-training partner for NVIDIA Nemotron. We post-train Nemotron models for customer use cases, de-risk mainline RL runs on our AC2 platform and training stack, and contribute aggregate workload statistics for inference benchmarking. This is how open models get better, and we're excited to keep working closely with NVIDIA AI.show more

Training large models with RL typically runs on a... fixed homogeneous cluster of the latest accelerators for numerical stability and faster throughput. This throws away inference optimization work done for other vendors and generations, limits the capacity you can utilize, and is unnecessarily expensive. So we trained a model on agentic coding tasks using fully asynchronous RL and mixed-precision, mixed-accelerator samplers. With off-policy correction, training stayed stable through convergence, on par with homogeneous samplers, at a lower cost.show more

Why Memory and GPUs Have to Be Co-Located “If... you want to do large-scale training like we do, you need the memory to be co-located with a large cluster of GPUs. I can’t just rent from Google, Microsoft, or even Base10 and run training at the scale I want because I need a gigantic memory card next to the GPUs, containing all my data and accessible by the entire cluster.” Cliff Weitzman Love to hear your thoughts on this Bryan Catanzaro Tri Dao Matei Zaharia Naveen Raoshow more

🚨 Every revolution starts with a wild idea. We... built a Robotics Gym, where anyone can train real robots powered by on-device AI. Welcome to the world’s first AI & Robotics Gym 👇show more

🚀 From smartphones to servers, each connected node fuels... the AI of tomorrow. 💡 In this new era, compute is the new oil and Hyra AI is building the pipeline with a decentralized network where your idle devices become powerful engines for AI. Explore more 👉show more

🚀 Meet Alaya_AI — the open & composable #Web3... infrastructure revolutionizing #AI data. 🦾 Powered by distributed communities, Alaya delivers intelligent sampling, ultra-precise auto-labeling, and high-quality datasets to train the next generation of AI models. - Swarm intelligence - Web3 incentives - RLHF + HITL - On-chain data 🔗CA ( $BNB ) : 0x5dBde81fcE337FF4bcaaEe4Ca3466C00aeCaE274 🌐 ➡️ The future of decentralized AI is being built right now. Who’s ready to contribute? 👀 #AlayaAI #Web3AI #DecentralizedAI #AIDatashow more

A 27-year-old guy from China bought 100 Mac minis... and turned his apartment into a server hub Instead of furniture, the space is filled with neat racks where a hundred Mac minis hum like a living organism, processing gigabytes of data in a continuous stream Many small AI startups in Asia cannot afford to train models on NVIDIA H100 servers: it is insanely expensive He provides his Mac minis for running and fine-tuning small, highly specialized models. He takes on the dirty work of data labeling and quality assurance by using his network of agents that check each other's work > His initial investment was $59,900 > He was able to recoup his costs in 3 months Right now, his average income is $25,000–$30,000 per month He is not a genius: he simply found a niche with high demand and created an offershow more

Announcing our $320M Series A at a $2.3B valuation,... led by Khosla Ventures, with General Catalyst, Eric Schmidt and Jeff Bezos. General Intuition is the frontier lab for acting in space and time. We build large action foundation models trained on billions of ground truth action-labeled gameplay clips from 17M monthly active users on Medal, and push the frontier of world models to generate infinite training environments.show more

☀️Very proud of the team for training 5.6! A... few of my highlights: - Great front-end aesthetics. E.g., I asked Sol to make a new blog for 5.6 with a celestial theme. - Much better CUA - Can work for much longer and is less lazy But it's not all great: - The writing quality has improved, but it's still bad in artifacts - Too many options with models/reasoning/fast mode/Cerebras/multi-agent We'll fix the above. What else should we improve? PS: belated post as there are so many other good models to train.show more

"It's on my todo to build my own stack... of apple minis and run Llama on my own little cluster on my desk." - Garry Tan Quick AI homelab tutorial: - Install EXO Labs on each mini (open source, steps in README) - Make sure all the minis are on the same WiFi network or for faster/more reliable results, connected over Thunderbolt/Ethernet. Devices will auto-discover each other and auto-shard the LLMs. - Now you can chat to your cluster with your preferred LLM GUI (the tiny corp tinychat ships with exo)show more

We found a new way to get language models... to reason. 🤯 No RL, no training, no verifiers, no prompting. ❌ With better sampling, base models can achieve single-shot reasoning on par with (or better than!) GRPO while avoiding its characteristic loss in generation diversity.show more

today's sota coding models train on hundreds of thousands... of coding simulation sandboxes all you need to do so is a sandbox where the agent solves problems (vLLM + openenv + opencode), a trainer (TRL) and a verifier to check solutions we just teach you how to do it for free (no need to thank it's ok)show more

1/ Gemini 2.5 is here, and it’s our most... intelligent AI model ever. Our first 2.5 model, Gemini 2.5 Pro Experimental is a state-of-the-art thinking model, leading in a wide range of benchmarks – with impressive improvements in enhanced reasoning and coding and now #1 on Arena by a significant margin. With a model this intelligent, we wanted to get it to people as quickly as possible. Find it on Google AI Studio and in the Google Gemini for Gemini Advanced users now – and in Vertex in the coming weeks. This is the start of a new era of thinking models – and we can’t wait to see where things go from here.show more

We are proud to introduce our new product, It... is a Virtual Machine specifically designed for Apple Silicon and AI Agents. This is our very first release, and it fully supports macOS, Windows, and Linux with GPU acceleration. - Liquid Glass, - GPU Accelerator. - MCP for AI Agent on your Mac or inside the VM.show more

Where physical AI meets the farm 🌱 Aigen is... using NVIDIA Cosmos and NVIDIA Jetson to build solar-powered robots for autonomous, chemical-free weed control, reducing reliance on herbicides and diesel-powered field passes. With NVIDIA technology, Aigen can: ✅ Train perception models with 99% synthetic data ✅ Reduce data collection and labelling to turn out effective models in hours instead of seasons Read the story:show more