Загрузка видео...

Не удалось загрузить видео

На главную

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.

exolabs's profile picture

EXO Labs

50,122 subscribers

347,553 просмотров • 1 год назад •via X (Twitter)

ОбразованиеНаука и технологии
LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

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.
0:18

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.

EXO Labs

719,005 просмотров • 1 год назад

M4 Mac Mini AI Cluster Uses EXO Labs with Thunderbolt 5 interconnect (80Gbps) to run LLMs distributed across 4 M4 Pro Mac Minis. The cluster is small (iPhone for reference). It’s running Nemotron 70B at 8 tok/sec and scales to Llama 405B (benchmarks soon).
0:19

M4 Mac Mini AI Cluster Uses EXO Labs with Thunderbolt 5 interconnect (80Gbps) to run LLMs distributed across 4 M4 Pro Mac Minis. The cluster is small (iPhone for reference). It’s running Nemotron 70B at 8 tok/sec and scales to Llama 405B (benchmarks soon).

Alex Cheema

3,515,852 просмотров • 1 год назад

if you’re confused about how large language models work this new animation from google deepmind should clear a lot of things up
0:18

if you’re confused about how large language models work this new animation from google deepmind should clear a lot of things up

anton 🇺🇸

874,031 просмотров • 2 лет назад

Preparing to run the new Llama 3.3 70B on my mac cluster. Downloading model shards onto 3 x M4 Pro Mac Mini and 1 x M3 Max MacBook Pro. AI cluster is connected by Gigabit ethernet switch with EXO Labs
0:18

Preparing to run the new Llama 3.3 70B on my mac cluster. Downloading model shards onto 3 x M4 Pro Mac Mini and 1 x M3 Max MacBook Pro. AI cluster is connected by Gigabit ethernet switch with EXO Labs

Alex Cheema - e/acc

181,559 просмотров • 1 год назад

Day 1: Benchmarks We ran 1000+ LLM benchmarks on real consumer devices. Data includes single-device and multi-device clusters with Tokens-Per-Second (TPS) and Time-To-First-Token (TTFT). Setups tested: 3x M4 Mac Mini cluster, iPhone 15 + S24, RTX4090 & more.
0:18

Day 1: Benchmarks We ran 1000+ LLM benchmarks on real consumer devices. Data includes single-device and multi-device clusters with Tokens-Per-Second (TPS) and Time-To-First-Token (TTFT). Setups tested: 3x M4 Mac Mini cluster, iPhone 15 + S24, RTX4090 & more.

EXO Labs

254,968 просмотров • 1 год назад

Presenting the world’s first autonomous, on-demand data network for specialized AI models. Turf DevNet is live on the Avalanche🔺 network🔺 Builders can now use contextual Data Feeds to train smarter AI 🧵
0:19

Presenting the world’s first autonomous, on-demand data network for specialized AI models. Turf DevNet is live on the Avalanche🔺 network🔺 Builders can now use contextual Data Feeds to train smarter AI 🧵

TURF

15,139 просмотров • 11 месяцев назад

Artificial Intelligence (#AI) models that process and interpret visual data are called large vision models. These models, like large language models, are trained on large amounts of data. #midjourney #dalle #stablediffusion (1/8) 🧵 ⤵️
0:10

Artificial Intelligence (#AI) models that process and interpret visual data are called large vision models. These models, like large language models, are trained on large amounts of data. #midjourney #dalle #stablediffusion (1/8) 🧵 ⤵️

Nuklai

13,614 просмотров • 2 лет назад

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 #GPUCluster
0:10

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 #GPUCluster

Itronix

19,906 просмотров • 1 год назад

m4 mac mini cluster with 5 nodes running exo
0:13

m4 mac mini cluster with 5 nodes running exo

Alex Ziskind

361,960 просмотров • 1 год назад

🌐 Experience the Power of Edge Computing with Hyra AI! We’re scaling at microsecond speed with Train-to-Earn: ⚡ 7.5B+ AI jobs completed 📱 2.2M+ active devices 💰 528M+ $HYRA rewards distributed Join a global movement of AI builders & earners. 🍎 iOS: 📱 Android:
0:10

🌐 Experience the Power of Edge Computing with Hyra AI! We’re scaling at microsecond speed with Train-to-Earn: ⚡ 7.5B+ AI jobs completed 📱 2.2M+ active devices 💰 528M+ $HYRA rewards distributed Join a global movement of AI builders & earners. 🍎 iOS: 📱 Android:

Hyra AI

22,236 просмотров • 11 месяцев назад

you can just ▲ ai gateway models ⬩ 1 api key to rule them all ⬩ AI SDK & @openai compat ⬩ globally redundant and distributed
0:15

you can just ▲ ai gateway models ⬩ 1 api key to rule them all ⬩ AI SDK & @openai compat ⬩ globally redundant and distributed

Guillermo Rauch

71,209 просмотров • 9 месяцев назад

The new M4 Mac Mini is sooo mini!!
0:10

The new M4 Mac Mini is sooo mini!!

iJustine

363,209 просмотров • 1 год назад

Improving tropical cyclone prediction can help communities better prepare & stay safe. Developed with Google DeepMind, we're launching Weather Lab to share our AI weather prediction models, starting with the new experimental cyclone prediction model. 🧵
0:13

Improving tropical cyclone prediction can help communities better prepare & stay safe. Developed with Google DeepMind, we're launching Weather Lab to share our AI weather prediction models, starting with the new experimental cyclone prediction model. 🧵

Google Research

68,852 просмотров • 11 месяцев назад

LLMs + Elixir: distributed tracing for targeted processes anywhere on the cluster with realtime UI. Idea to shipped, sub 30min:
0:10

LLMs + Elixir: distributed tracing for targeted processes anywhere on the cluster with realtime UI. Idea to shipped, sub 30min:

Chris McCord

12,014 просмотров • 4 месяцев назад

🚀 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 👉
0:11

🚀 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 👉

Hyra AI

12,373 просмотров • 11 месяцев назад

We are entering a new era of on-device automation. ✨ Watch Gemma 4 E4B navigate and drive an iOS simulator directly using Argent. Local models can handle complex interactions and software navigation autonomously.
0:17

We are entering a new era of on-device automation. ✨ Watch Gemma 4 E4B navigate and drive an iOS simulator directly using Argent. Local models can handle complex interactions and software navigation autonomously.

Google Gemma

471,894 просмотров • 15 дней назад

300+ devices. Millions of real-time signals. One decentralized weather oracle. We’re partnering with Nubila Network (mainnet arc) to bring hyperlocal, verifiable climate data to Nexus—fueling prediction markets, AI models, and on-chain protocols.
0:11

300+ devices. Millions of real-time signals. One decentralized weather oracle. We’re partnering with Nubila Network (mainnet arc) to bring hyperlocal, verifiable climate data to Nexus—fueling prediction markets, AI models, and on-chain protocols.

Nexus

17,911 просмотров • 8 месяцев назад

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.
0:11

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.

Aayush Karan

277,177 просмотров • 7 месяцев назад