Getting Started with Vertex AI Model Garden - Introduction to Foundation Models and MLOps

Explore a comprehensive technical session that introduces Vertex AI Model Garden, a platform hosting over 100 advanced LLMs and task-specific models including Palm 2 and open-source options like Llama 2 and Stable Diffusion. Learn to navigate the platform's key features, select appropriate models for specific use cases, and deploy them effectively through endpoints for application integration. Follow along with detailed demonstrations, including a walkthrough of Llama 2 on Google Cloud, and discover essential MLOps practices for model management. Master multi-task models, prompting techniques, model customization through tuning, and proper evaluation of generated outputs. Gain insights into enterprise data connectivity, generative AI evaluation services, and safety bias considerations. Access additional learning resources through the Generative AI for Developers Learning Path and engage with the Google Cloud Community for continued support and collaboration.

Syllabus

Google's first party foundation models
Introduction to Vertex AI Model Garden
Tune models to fit your use case
Demo walkthrough of Llama 2 on Google Cloud
Managing your models with MLOps
Multi-task models and prompting

Selecting pre-trained models and prompting
Customization with tuning and curated data
Managing new artifacts
Evaluating and monitoring generated output
Connecting to Enterprise data
Three building blocks to support the Gen AI evaluation service
Safety bias evaluation
Summarization of building your MLOps strategy for Generative AI
Q&A