Cerebras выкатила сервис с самым быстрым инференсом Llama 3.1 (~2k tk/s для 70B и ~1k для 405B).
Выглядит и правда круто, хотя понятно, что в ближайшем будущем все платформы будут предоставлять возможность практически мгновенного вывода — когда-то и картинки абсурдно долго загружались dialup-модемами.
🔗 cerebras.ai/inference
#LLM #Llama
Выглядит и правда круто, хотя понятно, что в ближайшем будущем все платформы будут предоставлять возможность практически мгновенного вывода — когда-то и картинки абсурдно долго загружались dialup-модемами.
🔗 cerebras.ai/inference
#LLM #Llama
Как доставляют причальные краны (в случае с первой пикчей — аж из Китая в Ливерпуль!)
🔗 investliverpool.com/news
🔗 investliverpool.com/news
🤔1
Malware can turn off webcam LED and record video, demonstrated on ThinkPad X230 🤦♂️
🔗 github.com/xairy/lights-out
🔗 github.com/xairy/lights-out
GitHub
GitHub - xairy/lights-out: Tools for controlling webcam LED on ThinkPad X230
Tools for controlling webcam LED on ThinkPad X230. Contribute to xairy/lights-out development by creating an account on GitHub.
Анонимный телеграм-канал
M4 Mac AI Coding Cluster Uses @exolabs to run LLMs (here Qwen 2.5 Coder 32B at 18 tok/sec) distributed across 4 M4 Mac Minis (Thunderbolt 5 80Gbps) and a MacBook Pro M4 Max. Local alternative to @cursor_ai 🔗 x.com/alexocheema #LLM #Hardware
YouTube
Finally a Mac for This Economy - Mac Mini M4
Get a free 15-day trial of Odoo’s all-in-one business solution and see how it can make your life easier! Check it out at https://www.odoo.com/r/5eLS
Apples M4 Mac Mini on paper seems to be too good of a deal, but are they hiding anything? We build a $600…
Apples M4 Mac Mini on paper seems to be too good of a deal, but are they hiding anything? We build a $600…
We present DeepSeek-V3, a strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token.
We pre-train DeepSeek-V3 on 14.8 trillion diverse and high-quality tokens, followed by Supervised Fine-Tuning and Reinforcement Learning stages to fully harness its capabilities.
Comprehensive evaluations reveal that DeepSeek-V3 outperforms other open-source models and achieves performance comparable to leading closed-source models. Despite its excellent performance, DeepSeek-V3 requires only 2.788M H800 GPU hours for its full training.
🔗 github.com/deepseek-ai
#LLM #GenAI #DeepSeek
We pre-train DeepSeek-V3 on 14.8 trillion diverse and high-quality tokens, followed by Supervised Fine-Tuning and Reinforcement Learning stages to fully harness its capabilities.
Comprehensive evaluations reveal that DeepSeek-V3 outperforms other open-source models and achieves performance comparable to leading closed-source models. Despite its excellent performance, DeepSeek-V3 requires only 2.788M H800 GPU hours for its full training.
🔗 github.com/deepseek-ai
#LLM #GenAI #DeepSeek
Самодельный «самый яркий игровой монитор» (участники гонки нит напряглись!):
https://www.youtube.com/watch?v=5FdDUrHl5RE
#Hardware
https://www.youtube.com/watch?v=5FdDUrHl5RE
#Hardware
YouTube
I made the World’s BRIGHTEST Gaming Monitor!
Check out Acronis Cyber Protect for businesses at http://go.acronis.com/LTT
Checkout NexiGo’s Anniversary Sale at: https://www.nexigo.com/pages/1st-anniversary-sale
Sure, TVs and Monitors have gotten brighter, but NOT BRIGHT ENOUGH! That’s why we embarked…
Checkout NexiGo’s Anniversary Sale at: https://www.nexigo.com/pages/1st-anniversary-sale
Sure, TVs and Monitors have gotten brighter, but NOT BRIGHT ENOUGH! That’s why we embarked…