Habana Labs

Bringing choice to #GenAI with performance, scalability, and efficiency. The new #IntelGaudi 3 AI accelerator is projected to average up to 1.3x faster inferencing vs. NVIDIA H200, running common Large Language Models. Learn more: https://intel.ly/3JcyM5h

We’re pleased to support MLCommons so our customers have transparent, reliable comparisons. The results show we’re raising the bar with competitive solutions for all your high-performance, high-efficiency deep learning inferencing needs — even on challenging LLMs. Read more about the results. https://intel.ly/3vvRbXA #IntelXeon #IntelGaudi #LLM #Intel

The latest MLPerf benchmark results are in! Check out how Intel Corporation continues to raise the bar for competitive #AI solutions. ⬇️

Intel® Gaudi®2 #AI accelerators support Open Federated Learning to enable collaborative training without sharing sensitive info. Check out the step-by-step guidance from Basem Barakat and satya ramji ainapurapu. https://intel.ly/3OvE9zv

"Gaudi®2 punches above its weight… matching the NVIDIA H100 in decoding latency" for inference of Llama 2 - 70B. Read the Databricks blog from Abhinav Venigalla and Daya Khudia to learn more. #AI #GenerativeAI https://intel.ly/3vozIQo

An exciting day at #NeurIPS2023! Intel Labs is demoing Intel Gaudi®2 AI Accelerators with its LDM3D model, the industry’s first #GenerativeAI model to deliver depth mapping. Learn more about LDM3D and how Gaudi accelerators are helping revolutionize content creation: https://bit.ly/414WgkT #GenerativeAI #AI #Content #3D

There’s still time to register! ⬇️ Join Intel Corporation’s #AI experts to learn more about how smaller #LLMs can help integrate #GenAI for increased performance, domain adaptation, private data integration, and more. See you on Thursday, December 7: https://intel.ly/3RSxoue #GenerativeAI #AI #Data #LLMs

ICYMI: Intel Corporation’s latest webinar covered key principles about domain-specific LLMs. Kevin Petrie and Ro Shah discussed how open-source LLMs can help accelerate all sorts of business functions, from customer service to document processing and more. Discover the next wave of #GenAI and watch the replay here: https://intel.ly/3RSxoue #GenerativeAI #AI #LLMs #Data

Building on the MLPerf MLCommons training results from June, new training benchmark results reinforce Intel Gaudi®2 as the only alternative for Nvidia’s H100 for #AI compute needs. Learn more about how Intel Gaudi®2 accelerators are delivering on performance: https://bit.ly/3Qryz1r #LLMs #Computing #AI #Data

Intel Gaudi®2 accelerators continue to raise the bar for #AI performance. 💪 We’re proud to announce the latest MLPerf MLCommons benchmark results: Intel Gaudi®2 demonstrated a 2x performance leap with the implementation of the FP8 data type on the v3.1 training GPT-3 benchmark. Read the full announcement: https://bit.ly/3Qryz1r #Training #LLMs #Data #AI