Amazon Web Services (AWS) announced at the AWS Summit in London that the Technology Innovation Institute (TII), a world-leading scientific research center in Abu Dhabi, trained its highest performing open source Falcon 40B model on AWS. announced that it is
AWS shares details about how we help customers across industries and how they can benefit from generative AI. To drive performance and cost efficiencies throughout the development process, customers including Stability AI, AI 21 Labs, Hugging Face, and LG AI leverage Amazon SageMaker, AWS’s end-to-end machine ML service. , builds, trains and deploys. LLMs.
Falcon 40B model uses AWS to meet specific customer needs
TII’s Falcon 40B model is now available to customers via Amazon SageMaker Jumpstart, a machine learning (ML) hub that offers pre-trained models. As a result, it becomes easier to build state-of-the-art generative AI applications using the best-performing open-source models.
Additionally, TII trained Falcon 40B, a 40 billion parameter LLM, on Amazon SageMaker. Ranked #1 on Hugging Face’s Open LLM Leaderboards as of June 2023.
Customers can deploy Falcon 40B on Amazon SageMaker JumpStart, giving them access to Falcon 40B’s accuracy and industry-leading performance without building their own model from scratch.
Open Source: Aims to allow broad access to advanced technical features.
Also, throughout the training process, AWS worked closely with TII to use SageMaker to enhance resilience, ensuring training runs smoothly and reducing interruptions that require developer attention. .
TII released the Falcon 40B model in May 2023 under the Apache 2.0 license. Since then, the Falcon 40B has outperformed similar contemporary models across a range of benchmarks, showing excellent performance without any special tweaks.
“The open-source release of Falcon 40B will enable organizations to leverage its outstanding capabilities and drive the advancement of AI-driven solutions. It is a stone and a demonstration of the deep scientific contribution of the UAE,” said Dr. Ebtesam Almazrouei, Executive Director of TII’s AI Cross Center Unit and Project Leader of the LLM Project.
“By making Falcon LLM open source, we aim to enable broad access to its advanced technical capabilities and empower researchers and organizations around the world. This includes contributing to further progress in the field of AI and advanced technology on the horizon, and promoting the use of advanced AI technology within UAE organizations and businesses.”
AWS Democratizes Access to Generative AI
The announcement also includes details on how TII built the highest performing Falcon 40B model on AWS and plans for Persistent Systems to enable Amazon CodeWhisperer for over 16,000 engineering organizations.
These announcements demonstrate how AWS is democratizing customer access to generative AI for a variety of needs, including transforming the way work is done with generative AI-powered tools like CodeWhisperer. is emphasized.
Persistent Systems builds and delivers industry applications to clients faster and more securely. Trained on billions of lines of code, Amazon CodeWhisperer enables persistence systems engineers to generate code based on natural language comments or existing code. This avoids time-consuming development tasks and, as a result, accelerates the creation of new solutions.
