Purple Hat’s cutting-edge cloud-based AI and machine learning platform features a comprehensive model registry with robust versioning and monitoring capabilities, sophisticated data drift detection and bias detection tools, and seamless LoRA fine-tuning functionality. According to Purple Hat, stronger safety measures are now available.

The upcoming release of Model 2.15 of Purple Hat OpenShift AI is scheduled for mid-November. Options highlighted within the expansive scope of the launch.

• A pioneering concept, the mannequin registry, currently in a cutting-edge preview state, offers a meticulously designed framework for sharing, modeling, deploying, and monitoring fashion, metadata, and mannequin assets with unparalleled ease?
• Detection of knowledge drift in deployed machine learning models, which involves monitoring shifts in the underlying knowledge distributions. This feature enables data scientists to identify instances where the deployed model’s performance significantly diverges from the training data it was originally trained on. Effective drift detection enables confirmation of mannequin reliability.
• Tools designed to help data scientists and artificial intelligence engineers gauge the credibility and impartiality of models, ensuring accurate insights and decision-making processes. The predictive tools, developed by the open-source TrustyAI group, not only monitor patterns but also scrutinize fairness in real-world implementations.

• Wonderful tuning with LoRA enables eco-friendly fine-tuning of large language models (LLMs), comparable to LLaMA-3. Organizations can thereby scale their AI workloads while reducing costs and conserving valuable resources.

• Designed to accelerate the provision of interfaces, a suite of streamlined tools is now available.
• Streamline your machine learning workflows with seamless integration of AMD GPUs using the AMD ROCm workbench – a comprehensive visual interface that empowers you to unlock the full potential of your AMD graphics processing units (GPUs) for model training and development.

Purple Hat OpenShift AI offers enhanced functionality for deploying and serving generative AI models, including the vLLM serving runtime for KServe, a cloud-native model inference platform? Additionally, support has been added for KServe Modelcars, enabling the storage and access of various model variations through Open Container Initiative (OCI) repositories. Moreover, personal and public route choices enable organisations to enhance the security posture of a model by selectively directing it to internal endpoints as needed.