Skip to main content

Introduction to AI Fine-Tuning

Just like you can tune a car to suit your preferences, you can also fine-tune AI to match your work needs.

By tuning AI, you can create a specialized AI for a specific field or a unique AI that understands dialects instead of standard language.

This process of further training a pre-trained AI to adapt it to specific tasks or objectives is known as Fine-Tuning.

Fine-tuning enables you to instill identity and personality into AI, transforming it into an expert assistant well-versed in specific areas.

This curriculum introduces foundational knowledge about what it means for AI to learn, what occurs inside AI models during the learning process, and overall basic understanding of artificial intelligence.

Following this, the focus is on experiencing the actual fine-tuning process and retrained AI models to understand what can be accomplished through fine-tuning and the processes involved.

We hope Introduction to AI Fine-Tuning significantly enhances your understanding of AI.


Curriculum Outline​

The first course, Introduction to AI & Fine-Tuning Basics, introduces the fundamental concepts and principles of AI learning and fine-tuning. It's designed to provide basic knowledge of AI even to those who know nothing about programming.

The second course, Experiencing Fine-Tuning, builds on the first course by allowing you to directly experience various fine-tuned AI models and fine-tuning processes.


1. Introduction to AI & Fine-Tuning Basics​

Introduces the fundamental concepts and principles of prompt engineering.


Chapter 1: Mastering the Basics​

  1. Our AI has changed! Creating Custom AI with Fine-Tuning

  2. Fine-Tuning vs. Prompt Engineering: What's the Difference?

  3. What are the Benefits of Fine-Tuning?

  4. How Generative AI Understands Prompts

  5. Fine-Tuning Models Specialized in Organizing Tabular Data

  6. Glossary for AI Beginners

  7. What is Deep Learning?

  8. How do Machine Learning and Deep Learning differ?

  9. What Does it Mean for AI to 'Learn'?

  10. How is Fine-Tuning Applied?


Chapter 2: Preparing Training Data​

  1. Data Formats Used to Train AI

  2. Deep Dive into JSON Data Format

  3. The Fuel for Fine-Tuning, What is JSONL?

  4. JSONL Format for Fine-Tuning

  5. Preparation for Training AI, Dataset

  6. Types of Datasets Used for AI Training

  7. How to Split Datasets

  8. Preprocessing: Making Data Easier to Digest

  9. What Data Formats Do Other AI Models Use?

  10. Creating JSONL Datasets

  11. Comparing Responses: General vs. Fine-Tuned Models


Chapter 3: Deciding on Training Methods​

  1. The Gap Between Ideal and Reality, Loss Function

  2. The Gradient of Loss Function

  3. Hyperparameters That Determine How to Train AI

  4. Remember These Three: Learning Rate, Batch Size, Epochs

  5. The Rate of Learning: Learning Rate

  6. The Scale of Data Trained at Once: Batch Size

  7. The Number of Times Training is Repeated: Number of Epochs

  8. Comparing Hyperparameters of Fine-Tuned Models

  9. Summary of Core Class Content




2. Experiencing Fine-Tuning​

Focuses on how to apply prompt engineering to real business scenarios.


Chapter 1: Learning While Trying Fine-Tuning​

  1. Converse with a Scholar from the Colonial Era

  2. What is a Base Model?

  3. What Process Does Fine-Tuning Complete?

  4. Creating a Model Using a Middle-Aged Woman’s Unique Tone

  5. Designing JSONL Datasets for Fine-Tuning

  6. Data Augmentation with AI

  7. Validating the Fine-Tuning Dataset

  8. Adjusting Hyperparameters

  9. Initiate Fine-Tuning

  10. Interacting with Fine-Tuned Models

  11. Comparing Base Models and Fine-Tuned Models


Chapter 2: What to Check After Training?​

  1. Evaluating Fine-Tuned Models

  2. A Closer Look at Overfitting

  3. Understanding Underfitting

  4. What is Training Instability?

  5. What is Imbalanced Data?

  6. What is Transfer Learning?

  7. Reference AI Training Code Example - Part 1

  8. Reference AI Training Code Example - Part 2

  9. How to Deploy Fine-Tuning Models?


Chapter 3: Fine-Tuning Playground​

  1. Creating an MBTI Dataset

  2. Setting Hyperparameters

  3. Initiate MBTI Fine-Tuning

  4. Testing Fine-Tuned Models

  5. Freely Interacting with Different MBTI Type Models

  6. Resources and Communities for Fine-Tuning Learning

  7. Comprehensive Recap of Fine-Tuning

  8. Epilogue


I hope Clear Guide to Prompt Engineering serves as a great aid in your work and research.