<img height="1" width="1" style="display:none;" alt="" src="https://px.ads.linkedin.com/collect/?pid=2552578&amp;fmt=gif">
footer-logo-2

Meet with Deci
at GTC 2024

March 17-21, 2024
San Jose, California

We’re exhibiting at the 2024 GTC AI conference.

Whether you are looking to achieve real-time performance, reduce model size, or increase throughput, Deci's NAS-based model optimization can help you deliver seamless inference in any environment. 

Book a meeting with Deci’s experts to learn how Deci’s Automated Neural Architecture Construction Engine can empower your team to build, optimize, and deploy highly accurate and efficient models for your deep learning and generative AI applications.

See you at booth #1501!

.

Catch Our Session on Gen AI Cost-Efficiency

Tuesday, Mar 19 | 4:30 PM - 4:55 PM PDT

Can High Performance also be Cost-Efficient when it Comes to Generative AI?

Large language models present a unique set of challenges when implemented in production environments. Their extensive size and complexity result in high operational costs and reduced inference speed, especially when scaling up. Traditional optimization techniques, tailor-made for smaller models, fall short for these models.

This talk illuminates the path toward efficient model design with a spotlight on DeciCoder, an open-source code generation LLM. DeciCoder was generated with Neural Architecture Search and designed to maximize performance on the NVIDIA A10 GPU. We'll explore the intricate methodologies that define the design and training phases of DeciCoder and demonstrate the transformative potential of neural architecture search for generative resource-efficient LLMs. Gain valuable insights to enhance your research, development, and deployment strategies.

Can you truly have your high-performance GenAI and deploy it cost-effectively, too? Join us to find out.

 

WATCH THE SESSION →

YG

Yonatan Geifman

CEO & Co-Founder, Deci

Yonatan Geifman is the CEO and Co-Founder of Deci. Before co-founding Deci, Yonatan was a member of Google AI’s MorphNet team. He holds a PhD in Computer Science from the Technion-Israel Institute of Technology and a B.Sc. and M.Sc. in Mathematics and Computer Science from Ben-Gurion University in Israel.

His research focused on making Deep Neural Networks (DNNs) more applicable for mission-critical tasks. It has been published and presented at leading global conferences including the Conference on Neural Information Processing Systems (NeurIPS) and International Conference on Machine Learning (ICML).

Don't Miss the Excitement at Booth #1501!

Explore Our Generative AI Models and Inference SDK

Join a live prompting challenge using our DeciLM + RAG demo.

Get to know more about DeciCoder, DeciDiffusion, DeciLM, and Infery LLM, and why you'll love them:

  • Efficient Model Design: Our generative models deliver high throughput and accuracy with low memory footprint and latency.
  • Supercharged Performance: Infery LLM ensures your AI runs smoother and faster, maximizing efficiency and dramatically reducing your cloud costs.

 

 

See Deci's Multimodal CV Model at the Edge

See a live stream inference of a multimodal model on a NVIDIA Jetson Orin Nano, and learn how Deci's foundation models:

  • Runs at unparalleled accuracy and speed, outperforming other well-known models. 
  • Is fully compatible with high-performance inference engines like NVIDIA® TensorRT™ and supports INT8 quantization.
  • Leverages cutting-edge techniques, such as attention mechanisms, quantization aware blocks, and reparametrization at inference time.
  • Is easy to fine-tune to achieve SOTA results using Google Colab Notebook with the SuperGradients open-source library.

 

 

Talk with Deci's Experts 1-on-1

Get a free consultation session and learn how other leading AI teams are collaborating with Deci to solve AI design and deployment challenges

DaveS

David Stein
Director of Strategic Partnerships
LinkedIn →

Elana Krasner

Elana Krasner
Senior Product
Marketing Manager
LinkedIn →

Aviad Simon

Avid Simon
 VP Operations and GM, US
LinkedIn →

Assaf-Katan

Assaf Katan
Chief Business Officer
LinkedIn →

Drew Allen

Drew Allen
Sr. Enterprise Account Executive
LinkedIn →

 

evs-meet-experts

 

AutoNAC (Neural Architecture Search Engine)

Automatically find the architecture that delivers the highest accuracy for your specific speed, size, and inference hardware targets

Tailored for Your Task & Performance Goals

Run a multi-objective search to generate an architecture optimized for several parameters (accuracy, throughput, latency, model size, and memory footprint)

Hardware Aware Neural Architecture Search

Take hardware constraints into account to unlock optimal efficiency that hit your performance targets

Outperforms any SOTA Computer Vision Model

Deliver the best accuracy and speed for your specific use case that outperform SOTA open-source neural networks

 

 

See You at Booth #1501!

1501

Locate our booth at the MLOps & LLMOps Platforms pavilion!

.