Presentation
BYOD Lab: Advanced AI Insights Leveraging Hugging Face Large Language Models as Part of a Multi-Model Approach on IBM Z and LinuxONE
DescriptionPredictive AI and Large Language Model deployments can be combined in a hybrid approach called multiple model to deliver more accurate results and more insightful predictions than any single model can accomplish alone, without compromising speed. This lab will guide you through the principles and practical implementation of multiple model AI on the IBM z17 platform, demonstrating how these innovations can dramatically improve prediction accuracy and business value. We will demonstrate using the Linux on Z technologies found in the AI Toolkit for IBM Z and LinuxONE.
1. We will run a pre-built hugging face container.
2. We will explore the content of a notebook that implements and deploys an encoder LLM BERT based model that is designed to extract key insights from the unstructured data.
3. We will explore the content of a notebook that builds and trains a Random Forest based model, which is intended to provide the final predictions using the combined data.
4. We will run the step-by-step guided inferencing notebook using PyTorch from AI toolkit for IBM
1. We will run a pre-built hugging face container.
2. We will explore the content of a notebook that implements and deploys an encoder LLM BERT based model that is designed to extract key insights from the unstructured data.
3. We will explore the content of a notebook that builds and trains a Random Forest based model, which is intended to provide the final predictions using the combined data.
4. We will run the step-by-step guided inferencing notebook using PyTorch from AI toolkit for IBM
Event Type
Technical Session
TimeTuesday, February 243:45pm - 4:45pm EST
LocationSTE Lab
Machine Learning/AI
New and Innovative Technologies
Hands-on Lab
All Audiences