LLM (Large Language Model) outputs and behaviors are non-deterministic. This course will teach you fine-tuning and LLM customization in order to better match user preferences and expectations.

What you'll learn

LLMs (Large Language Models) have general, limited knowledge and training data, and they may produce irrelevant, unsafe, or unhelpful outputs. In this course, Fine-tuning and Customizing LLMs, you'll learn to pre-train, fine-tune and reinforce the LLM learning to maximize their capabilities and align with human preferences and expectations.

First, you'll explore the training techniques to pre-train the models with supervised fine-tuning , prompt engineering, and reinforcement learning with human feedback. Next, you'll explore core training techniques, including supervised fine-tuning, prompt engineering, and reinforcement learning with human feedback (RLHF) to adapt pre-trained models for specific tasks and performance goals. Finally, you'll discover how to evaluate and optimize model performance by applying methods like parameter-efficient fine-tuning (PEFT), transfer learning, and evaluation metrics to improve model reliability and reduce bias.

When you're finished with this course, you'll have the skills and knowledge of fine-tuning and customizing LLMs that is needed to deliver accurate, context-aware, and human-aligned results across a wide range of AI-driven applications.

Homepage