Red Hat has announced the release of a new edition of the RHEL distribution – Red Hat Enterprise Linux AI (RHEL AI), specifically tailored for machine learning tasks and designed to simplify the creation of server solutions utilizing large language models.
The distribution includes a collection of tools and frameworks for machine learning, drivers for various hardware accelerators from AMD, Intel, and NVIDIA, and components to leverage the capabilities of Dell, Cisco, HPE, Lenovo, and SuperMicro servers optimized for AI systems. The builds are available through the Red Hat Customer Portal in options for direct installation on servers and for deployment in cloud systems AWS and IBM Cloud.
The platform can be used to develop enterprise AI applications, implement content generation services, create conversational systems, and integrate into virtual assistant applications that support skills such as answering questions in natural language, solving mathematical problems, generating meaningful text on a given topic, summarizing content, correcting text errors, performing rewriting, assisting in code writing in various programming languages, and generating letters and documents based on templates.
RHEL AI integrates components for the creation, testing, and execution of machine learning systems based on the large language model Granite, open-sourced by IBM under the Apache 2.0 license, capable of considering up to 4,000 tokens when generating text and encompassing 7 billion parameters. The open toolkit InstructLab is integrated into the distribution for interaction with the Granite model, supporting the LAB (Large-scale Alignment for chatBots) methodology for customization and optimization of models, as well as for adding additional knowledge and implementing new skills in pre-trained models.
Source: opennet.ru
