Eitan Medina

Formerly, Chief Operations Officer at @HabanaLabs, an @Intel company.

[Eitan Medina @emed4090]

https://t.co/wokkyP5As3

[Eitan Medina @emed4090]

https://t.co/vvI9clSno7

[Eitan Medina @emed4090]

Habana® Gaudi2—THE viable alternative to Nvidia GPUs for training #LLMs.

Read the results ➡️ https://t.co/r3hBDtqI1z

[Habana Labs @HabanaLabs]

Habana® Gaudi2 is one of only 2 semiconductor solutions to submit GPT-3 results on @MLCommons @MLPerf Training 3.0 benchmark.

These results validate Gaudi2 as THE viable alternative to H100 for training #LLMs.

Read about our results ⬇️
https://t.co/K6dwrtUHZd

[Habana Labs @HabanaLabs]

The Habana® team is collaborating with @GenesisCloud_ to deliver a new class of cloud instances and enable high-performance, high-efficiency #DeepLearning training and inference in the cloud ☁️

Learn about how we're helping customers with more for less ➡️ https://t.co/tblQCSLzE5

[Habana Labs @HabanaLabs]

Check out @awscloud blog to learn how DeepSpeed optimization for @PyTorch can help you efficiently scale large language models on first-gen Gaudi-based Amazon EC2 DL1 instances. 🧠

Check out the full blog ➡️ https://t.co/vdfObTb5BH

[Habana Labs @HabanaLabs]

We'll be there! Come see us at @intel booth #A105 to connect and check out our demos showing inference on large language models, from 12B to 176B parameters...and more!

#ISC23

[Habana Labs @HabanaLabs]

With the power of deep learning-optimized Habana Gaudi® accelerators, @BCG and @Intel created a custom #GenerativeAI model for hybrid compute environments, secure training, and inference.

Read more ⤵️

[Habana Labs @HabanaLabs]

🚨This just in...

With the recent drop of our 1.9 SynpaseAI software, Habana® Gaudi2 performance on 176B BLOOMZ went from 20% faster than A100 to 30%! 📈📈

Read the @huggingface blog on our faster inference performance on the world's largest #LLM👇
https://t.co/biZenvNNv2

[Habana Labs @HabanaLabs]

Check out @intel's @tonymongkolsmai's latest podcast with Ke Ding and @huggingface's @julsimon all about Gaudi®2 and the challenges and solutions around optimizing the latest #AI models, #ChatGPT, and more.

Listen in w/ @ApplePodcasts ➡️ https://t.co/k7sX5SP4Nz

[Habana Labs @HabanaLabs]

Did you know Habana has resources so you can level up 📈 your #DeepLearning skills?

Check out our tutorials like these on transfer learning, large language models, and getting started on Gaudi/Gaudi2. ⬇️
https://t.co/wPXq2D4k1f

[Habana Labs @HabanaLabs]

The Habana team advances Gaudi and Gaudi2 usability with the release of SynapseAI® Software version 1.9.0.

Learn more about the upgraded versions of several libraries, including @LightningAI on our #developer blog ➡️ https://t.co/SRPiakLwVb

[Habana Labs @HabanaLabs]

We're excited to collaborate w/ @EquusCS on their lab-as-a-service featuring Gaudi®2 #DeepLearning training & inference solutions.

Equus' LaaS offers full-stack #AI solutions based on Gaudi2 for accelerating workloads on-premise or in the cloud.☁️

More➡️https://t.co/Mz044RAPUa

[Habana Labs @HabanaLabs]

The Habana team is pleased to team with Colossal-AI to bring Gaudi®'s high-performance, high-efficiency #DeepLearning acceleration to protein folding prediction.

Learn more via @HPCAITech ➡️ https://t.co/SAs21VpXHI

[Habana Labs @HabanaLabs]

On our latest Developer blog, learn how to easily deploy multi-billion parameter #LLMs on Habana Gaudi2® and get a view into the @huggingface's performance evaluation of Gaudi2 and A100 on BLOOMZ.

Easily put high-performance #LLMs to work with Gaudi2 ➡️ https://t.co/DW2lpwcTXt

[VentureBeat @VentureBeat]

Hugging Face shows how @intel is growing the use of its hardware in the #generativeAI space, competing with @nvidia but with elements of cooperation as well. @huggingface https://t.co/OyGulsoA20

[Habana Labs @HabanaLabs]

With @huggingface, Habana's Gaudi2® runs inference faster on multi-billion parameter #LLMs.

[Habana Labs @HabanaLabs]

Watch research scientists with @intel's #AI team, Phillip Howard & Anahita Bhiwandiwalla, put Gaudi2 and 176-billion-parameter BLOOMZ to the test.

Discover the latest @intelnews on how you can easily put #LLM like BLOOMZ to work on Habana's Gaudi2® ➡️ https://t.co/aYJDn9jpK4

[Habana Labs @HabanaLabs]

Habana's Gaudi2® outperforms A100 on BLOOMZ, the world's largest #LLM, with faster inference performance. Read the full @huggingface blog to learn about this achievement!👇👇
https://t.co/biZenvNNv2

[Tony Mongkolsmai @tonymongkolsmai]

Another from ⁦@huggingface⁩ and Intel AI folks. Want to know why working at ⁦@HabanaLabs⁩ rocks? Cause they build fast AI hardware. #gaudi2 is cool tech folks 😊👍

Fast Inference on Large Language Models: BLOOMZ on Habana Gaudi2 Accelerator https://t.co/9f22O64WPo

[Habana Labs @HabanaLabs]

#Developers: want to learn how to get started with inference and running deep learning models on the Gaudi® processor?

Check out our on-demand webinar ➡️ https://t.co/rz85ZNMI9O