Red Hat has announced a collaboration with Lenovo to deliver Red Hat Enterprise Linux AI (RHEL AI) on Lenovo ThinkSystem SR675 V3 servers.
Earlier this year Red Hat South Africa outlined its AI strategy aimed at easing customers' AI adoption.
The open source company, and an IBM subsidiary, says GenAI offers organisations a chance to accelerate innovation, but AI workloads need high-performance servers and flexible software to handle intensive computing and streamline model training and fine-tuning.
Joe Fernandes, GM of Red Hat's AI business unit, says the collaboration will help customers run AI workloads across datacenters, clouds, and the edge. Lenovo's Brian Connors, GM of ESMB & AI business segment, says using RHEL AI on Lenovo servers will speed up AI workflows and time-to-value.
RHEL AI foundation
RHEL AI is a platform that combines IBM Research's Granite family of open source large language models (LLMs) with InstructLab's alignment tools, offering a community-driven approach to model development.
Rad Hat says the platform is now validated to support Lenovo's SR675 V3 servers and will be factory pre-loaded for customers to develop, test and run AI and GenAI models on a trusted foundation model platform, built on an AI-optimised operating system.
The solution comes as an optimised, bootable Red Hat Enterprise Linux image for hybrid cloud portability. It is part of Red Hat OpenShift AI, a hybrid cloud MLOps platform that scales models and InstructLab across distributed clusters. Lenovo Consulting Services offers support for optimising and implementing Red Hat on its servers.
Hybrid AI
At the Lenovo Tech World ’24 event in Seattle this week, smarter AI for all featured as the dominant theme.
Lenovo CEO Yuanqing Yang reiterated the Chinese company's view that hybrid AI is the future, with enterprise and public cloud models coexisting. He noted growing demand for private cloud AI training and inference using enterprise data.
Share