Beginner's Guide to Building with Open Source LLM

Join this session for a comprehensive introduction to building applications with open-source Large Language Models (LLMs). We will cover essential concepts, including the intuition behind LLMs, tokenizers, encoders, and decoders, as well as the distinctions between encoder-decoder models.

Topics include quantization techniques, fine-tuning methods like LoRA and adapters, and alignment strategies including RHLF, DPO, PPO, and ORPO. Additionally, we'll discuss Retrieval-Augmented Generation (RAG) and key metrics for evaluating LLMs. We'll also address the drawbacks of using open-source LLMs and provide practical solutions to overcome these challenges. This session is designed to equip beginners with the knowledge and tools to effectively build applications using open-source LLMs.