Red Hat recently released the latest version 1.3 of its enterprise-level Linux AI platform, RHEL AI. This version adds support for the IBM Granite large language model (LLM) and previews support for the Intel Gaudi3 accelerator. This update provides more opportunities for Red Hat's service partners and system integrators, helping businesses to explore and integrate AI applications more effectively.
Joe Fernandez, Vice President and General Manager of Red Hat AI, stated that service partners and system integrators play a crucial role in helping companies implement various application cases. He noted that the complexity of generative AI enables these partners to offer more cost-effective solutions to customers. Red Hat aims to reduce costs by using smaller models, simplifying the integration of models with customer data and use cases, and providing flexibility for deploying these models in hybrid environments.
RHEL AI version 1.3 is specifically designed for developing, testing, and running generative AI models for enterprise applications. This update includes support for the IBM open-source licensed Granite large language model and utilizes open-source technology for data preparation. Users can create deployable and bootable RHEL images for individual server deployment in hybrid clouds by aligning projects with the InstructLab model developed in collaboration with IBM.
The new version supports English language use cases for Granite 3.08b and offers a developer preview, allowing users to experience the model's capabilities in non-English languages, code generation, and function calling. Future versions will fully support these features. Additionally, RHEL AI supports the Docling open-source project from IBM Research, which can convert common document formats into Markdown, JSON, and other formats for generative AI applications and training. The new version features context-aware chunking capabilities that consider the structure and semantic elements of documents, aimed at improving the response quality of generative AI.
This release also adds technical preview support for the Intel Gaudi3 accelerator, allowing users to process multiple requests in parallel during real-time processing and dynamically adjust LLM parameters during processing. Future RHEL AI versions will further support more document formats from Docling, integrate enhanced retrieval generation pipelines, and include InstructLab knowledge tuning features.
Highlights:
🔹 RHEL AI version 1.3 adds support for the IBM Granite large language model, enhancing enterprise application capabilities.
🔹 The new version previews support for the Intel Gaudi3 accelerator, improving real-time processing capabilities.
🔹 Red Hat is committed to reducing the cost and complexity of AI models, promoting flexible hybrid cloud deployment.
