Join us for an enlightening session at Communitech on May 21st, 2024, from 8:00 AM to 9:30 AM, as we delve into the practical aspects of building Retrieval Augmented Generation (RAG) based solutions. Our guest speaker Daniel Manary will share his firsthand experiences and insights from his latest project.
Daniel's project involved deploying a quantized Mistal 7B Instruct model to process and respond to queries from a document chunked and stored in a vector database.
This session will cover essential topics such as the trade-offs between model size and quality, including quantization effects; choices in model selection; techniques in model prompting and prompt analysis; strategies for output analysis; and methods for data parsing and storage.
We will also introduce the critical aspect of output validation, enriching our understanding of applied RAG frameworks.
Whether you are a founder, developer, enthusiast, or simply curious about Generative AI, this session promises a comprehensive overview of building and implementing RAG-based solutions. Don’t miss the chance to gain valuable insights from a practitioner’s perspective and engage in discussions on the evolving applications of Generative AI in real-world scenarios.