Red Hat introduces updates for Red Hat OpenShift AI and Red Hat Enterprise Linux AI. The adjustments aim to help companies deploy AI models more efficiently in hybrid cloud environments.
New Features in OpenShift AI and RHEL AI
The latest version of Red Hat OpenShift AI, version 2.18, includes multiple improvements focused on ease of use and performance. OpenShift AI provides a platform for managing AI workloads, with support for MLOps and LLMOps.
New features include distributed model serving via the vLLM inference server. This enables faster processing by distributing AI models across multiple GPUs. Furthermore, InstructLab in combination with OpenShift AI data science pipelines makes it easier to fine-tune models to company-specific data.
read also
Red Hat Renews AI Portfolio for Hybrid Cloud Environments
AI guardrails ensure that models are less likely to generate unwanted results, such as inappropriate language or sensitive information. The new lm-eval allows organizations to test their AI models on various tasks, including text comprehension and logical reasoning.
Red Hat Enterprise Linux AI (RHEL AI) is also being updated to version 1.4. This platform now includes support for the Granite 3.1 8B model. This model works in multiple languages and offers a larger context window. The release also includes a new visual interface for data analysis and a benchmark tool (DK-bench) to compare AI models.
InstructLab on IBM Cloud and Training
Additionally, Red Hat AI InstructLab will be available on IBM Cloud. This environment offers companies a way to train and deploy AI models more securely and easily. Organizations maintain control over their data in this process.
Finally, Red Hat is launching free AI Foundations training. These are available to customers and provide an entry path for both beginners and experienced users who want to apply AI in their organization.
The latest versions of OpenShift AI and RHEL AI are immediately available. Red Hat AI InstructLab on IBM Cloud will follow later this spring.