From InstructLab and Podman AI Lab to OpenShift AI: Simplifying LLM workflows
summary
A complete local-to-cluster workflow showing how InstructLab fine-tunes Granite-based models (producing a DavcomCZ V2), how Podman AI Lab runs that model in a chatbot recipe on a 24 GB CPU laptop, and how the same container and GGUF/quantized model are pushed to OpenShift AI with MinIO S3 and an Elastic vector DB for RAG. GPU runs take ~15 min synthetic data + ~45 min fine-tune; CPU runs take ~1.5 hr + 2 days. The chatbot and RAG deployment work and return the conference dates.
AI-written summary · about
chapters (5)
- 0:00LLM workflow integration
- 7:45Training with Instruct Lab
- 15:51Testing with Podman AI Lab
- 21:10Deploying to OpenShift AI
- 28:15Q&A
related talks
37mAI Lab Recipes: Cooking up AI Applications on your Laptop
26mWelcome to the AI Jungle! Now What?
29mPodman AI Lab: For developers to build AI Applications with LLMs running locally 
47mRamaLama: Making working with AI Models Boring
01h 01mScaling AI with Java: Building Production-Ready LLM Applications
45mWelcome to the AI jungle! Now what?
28mProduction-Ready LLMs on Kubernetes: Patterns, Pitfalls, and Performa...
01h 36mTutorial: From Planni...