Unlock robust AI governance and streamline your MLOps workflow! Just like a pedigree helps track a puppy’s traits, the Model Registry in Red Hat OpenShift AI provides essential information and tracking for your AI models.
As AI infusion grows, well-structured AI governance is crucial to track deployed models, establish provenance, and bake in AI safety practices. The Model Registry acts as a single pane of glass, helping you track models throughout the entire lifecycle, from inception (inner loop) to production deployment (outer loop).
See the Model Registry in action within the OpenShift AI dashboard.
Time stamps:
• 00:6 – Introduction & AI Governance
• 00:35 – Model Lifecycle Overview
• 01:10 – Registering Models & Versions
• 01:50 – Adding Labels & Properties
• 04:15 – Deploying Models
• 05:45 – Undeploying & Archiving
• 06:30 – GenAI & LLM Ops Support
• 06:50 – Using the Model Catalog
• 09:15 – Managing Fine-Tuned Models
Matteo Mortari, a Principal software engineer in Red Hat’s AI division, discusses and demonstrates how to register models and versions, add labels and properties, manage deployment guardrails, undeploy and archive versions, and use standard labels for AB testing. It’s vital for tracking provenance.
For Generative AI (GenAI) and LLM ops, the Model Catalog surfaces vetted models. You can register these foundation models for fine-tuning and register your customized versions for deployment in your AI-powered applications.
Key benefits: Centralized tracking, Provenance, AI Governance, Lifecycle Management, Version Control, Ease of deployment, GenAI/LLM Support.
Ready to govern your AI models? Watch the demo!
Learn more: https://red.ht/AI
