LLM Engineering & Generative AI · LLM Engineering & Generative AI

LLM Finetuning

This track covers the practical skills for adapting pretrained large language models to your own data. You'll learn data preparation, parameter-efficient methods like LoRA and QLoRA, training loops with Hugging Face Transformers, and evaluation strategies. Ideal for ML engineers and developers wanting to customize LLMs for specific tasks, it ends with deploying a fine-tuned model responsibly.

Sponsored

Lesson outline

  1. 1
    What Is LLM Finetuning?

    Understand what LLM finetuning means and why it matters. This lesson explains the core concepts, walks through a practical example, and prepares you for hands-on fine-tuning in the next steps.

  2. 2
    Choose a Model for Use Case

    Learn to choose a model for your use case in this LLM Finetuning tutorial. Step-by-step guidance on evaluating models, comparing options, and practical tips for your finetuning projects.

  3. 3
    Set Up Python for Finetuning

    Set up a Python environment for finetuning with this hands-on LLM Finetuning tutorial. Learn the essentials, avoid common pitfalls, and get ready for the next lesson.

  4. 4
    Install HF Transformers & Datasets

    Install Hugging Face Transformers and datasets — LLM Finetuning lesson. Hands-on setup for fine-tuning pipelines.

  5. 5
    Causal LM Architectures

    Explore model architectures for causal LM in this LLM Finetuning tutorial. Understand key concepts, apply them in hands-on exercises, and get ready for the next lesson.

  6. 6
    Preparing Text Data for Finetuning

    Learn how to prepare text data for supervised LLM finetuning: formatting, tokenization, and quality checks, with a hands-on exercise.

  7. 7
    Build a Dataset with the Datasets Library

    In this LLM Finetuning lesson, you'll learn to create a dataset using the Hugging Face datasets library—covering loading, preprocessing, and saving your data for fine-tuning. Hands-on steps and troubleshooting included.

  8. 8
    Tokenize Your Data for Training

    Master tokenization for LLM fine-tuning: learn how to convert raw text into model-ready tokens, handle padding and truncation, and avoid common pitfalls. Includes a hands-on exercise.

  9. 9
    Attention Masks & Padding

    Understand attention masks and padding in LLM finetuning with this hands-on tutorial. Learn how padding ensures uniform batch lengths and attention masks prevent the model from attending to padding tokens. Includes step-by-step examples, troubleshooting, and next steps.

  10. 10
    Split Data into Train and Validation Sets

    Split data into train and validation sets — LLM Finetuning.

  11. 11
    Load a Pretrained Model with AutoModelForCausalLM

    Load a pretrained model with AutoModelForCausalLM — LLM Finetuning.

  12. 12
    Run Your First Trainer Loop

    Learn to run your first training loop with Hugging Face Trainer in this hands-on LLM finetuning lesson. Set up training arguments, call trainer.train(), and troubleshoot common issues.

    LLM Finetuning — step-by-step tutorials

    What you will find here

    This track walks through llm finetuning in order — each lesson is server-rendered HTML you can read without JavaScript. Follow the outline, then practice in the browser IDE when a lesson links to runnable code.

    Tutorials vs quizzes and code samples

    Tutorials teach in sequence. For quick checks use quizzes. For copy-paste snippets see code samples. For deeper reading browse articles.