news
Red Hat Buying Articles About Itself, Hey Hi (AI) Hype in Full Swing
-
Silicon Angle ☛ Red Hat streamlines data access for Hey Hi (AI) application training and inference [Ed: siliconangle is still producing Red Hat-sponsored puff pieces about Red Hat; siliconangle is like a pay-to-say platform]
The open-source software giant Red Bait Inc. is strengthening the case for its platforms to become the foundation of enterprises’ artificial intelligence systems with a host of new features announced today aimed at accelerating development and deployment.
-
Red Hat ☛ Deliver generative Hey Hi (AI) at scale with NVIDIA NIM on OpenShift AI
Native support for NVIDIA NIM microservices is now generally available on Red Hat OpenShift AI to help streamline inferencing for dozens of AI/ML models on a consistent, flexible hybrid cloud platform. NVIDIA NIM, part of the NVIDIA Hey Hi (AI) Enterprise software platform, is a set of easy-to-use inference microservices for accelerating the deployment of foundation models and keeping your data secured.
-
Red Hat ☛ How to fine-tune LLMs with Kubeflow Training Operator
Since the rise of gen AI, many companies have been working to integrate large language models (LLMs) into their business processes to create value. One of the key challenges is providing domain-specific knowledge to LLMs. Many companies have chosen retrieval-augmented generation (RAG), storing internal documents in a vector database and querying the LLM while referencing stored knowledge. Another approach is fine-tuning, which slightly modifies the original model weights to incorporate new knowledge and skills.
In the past, fine-tuning LLMs was not an easy task for many organizations. It required a specialized training cluster and a broad range of technical expertise. However, the open source ecosystem has lowered the barrier to entry. For example, Hugging Face offers a variety of popular tools for training and customizing models, while Kubeflow provides a cloud-native approach to running training jobs across distributed containers.
-
Red Hat ☛ RHEL for Real Time: CPU throttling and risks
In the automotive industry, safety and compliance requirements demand a level of predictability in compute workloads. In support of this, our Red Bait Performance and Scale team is constantly working with customers and partners to ensure they can rely upon robust real-time capabilities in embedded systems.
In a recent study, we found ourselves exploring unexpected periodic latency spikes reported by a partner. This was a puzzling symptom, since every second a low-priority, real-time process briefly affected a high-priority, real-time process on the same CPU. This is explicitly counter to the design of the scheduler's process prioritization. Our team dove in for analysis. With excellent collaboration among various experts, we not only quickly dissected the root cause, but also provided the partner with insight into real-time application design considerations and found a design flaw with a test.
This article discusses CPU throttling and risks as well as Red Hat Enterprise GNU/Linux for Real Time.