Red Hat, Inc., a global leader in open-source solutions, has announced the latest release of Red Hat Enterprise Linux AI (RHEL AI), an advanced platform designed to seamlessly develop, test, and run generative artificial intelligence (gen AI) models for enterprise applications. RHEL AI 1.3 introduces significant advancements in supporting the Granite large language model (LLM) family and incorporates innovations for data preparation. This release enhances hybrid cloud deployment options and features accelerated compute architecture, setting a strong foundation for enterprise-level AI applications.
- RHEL AI 1.3: Key Features and Advancements
- Granite 3.0 LLM Support: RHEL AI 1.3 extends its capabilities to support Granite 3.0 8b LLMs, enabling a broader range of use cases, including support for English and other languages, as well as code generation and function calling. Non-English language support and code generation are in developer preview, with more features expected in future releases.
- Optimized AI Models: Red Hat’s vision focuses on smaller, open-source-licensed models that can run efficiently across hybrid cloud environments, offering better performance and flexibility.
- Simplifying Data Preparation with Docling
- Docling Integration: Open-sourced by IBM Research, Docling is now part of RHEL AI 1.3, helping users convert document formats like PDFs into Markdown for easier data ingestion and model tuning. This feature significantly simplifies the process of preparing data for generative AI applications.
- Context-Aware Chunking: Docling’s context-aware chunking takes into account document structure and semantics, ensuring that the AI applications maintain coherence and contextually accurate responses.
- Hybrid Cloud and Hardware Architecture Support
- Broadening Gen AI Ecosystem: RHEL AI 1.3 supports various hardware architectures, including accelerators from NVIDIA, AMD, and Intel Gaudi 3 (in technology preview). This platform is optimized for hybrid cloud environments, supporting major cloud providers like AWS, Google Cloud, and Microsoft Azure, offering flexibility for different cloud setups.
- Chip Architecture and Accelerated Hardware: The release extends support for Red Hat partners’ accelerated hardware offerings, such as Dell PowerEdge R760xa servers and Lenovo ThinkSystem SR675 V3 servers.
- Enhancing Model Serving with Red Hat OpenShift AI
- Parallelized Model Serving: RHEL AI integrates with Red Hat OpenShift AI, enabling parallelized LLM serving across multiple nodes, which allows for real-time handling of multiple requests. This feature helps improve response times, customer satisfaction, and system scalability.
- Dynamic Model Parameters: OpenShift AI allows dynamic adjustments to LLM parameters, such as model sharding and quantization, optimizing the performance and footprint of models in real-time.
- RHEL AI in the Broader Red Hat AI Portfolio
- Seamless Integration with Red Hat OpenShift AI: RHEL AI, along with Red Hat OpenShift AI, forms the backbone of Red Hat AI, a portfolio designed to accelerate time-to-market and reduce the operational costs of deploying AI solutions across hybrid cloud environments. The integration of both solutions offers robust support for machine-learning operations (MLOps) and scale management.
The launch of RHEL AI 1.3 strengthens Red Hat’s position as a leader in the enterprise AI space. With support for the latest advancements in the Granite LLM family, seamless data preparation through Docling, and expanded hybrid cloud deployment capabilities, RHEL AI offers businesses the flexibility and power needed to integrate generative AI into their operations. As part of the broader Red Hat AI ecosystem, RHEL AI continues to drive innovation in AI deployment and management across hybrid cloud environments.
