Overview
Master Python classes and their application in AI development, focusing on LLM fine-tuning and Vision transformer implementation with KERAS3 functionality.
Syllabus
-
- Introduction to Python Object-Oriented Programming
-- Overview of Python classes and objects
-- Attributes and methods
-- Inheritance and polymorphism
- Advanced Python Class Features
-- Class and static methods
-- Property decorators
-- Dunder methods for operator overloading
- Python Classes for AI Development
-- Designing reusable classes for AI tasks
-- Understanding class hierarchies in AI models
-- Implementing model parameters as class attributes
- Introduction to KERAS3 and its Functional API
-- Overview of KERAS3 functionalities
-- Building models using the Functional API
-- Advantages of subclassing in KERAS3
- Subclassing in Large Language Models (LLMs)
-- Creating custom layers for LLMs
-- Fine-tuning pre-trained LLMs with subclassing
-- Case study: Implementing and fine-tuning a GPT-like model
- Fine-tuning Vision Transformers (ViT) using Python Classes
-- Overview of Vision Transformers architecture
-- Building and customizing ViT using KERAS3
-- Practical example: Fine-tuning a ViT on a custom dataset
- Best Practices and Optimization
-- Strategies for efficient model subclassing
-- Managing resources and optimizing performance
-- Debugging and testing custom AI model classes
- Capstone Project: Implementing a Custom AI Model
-- Define a project scope using Python classes
-- Design, implement, and fine-tune a custom AI model
-- Present and evaluate your AI model
- Conclusion and Further Resources
-- Review key concepts
-- Recommended readings and resources for further learning
Taught by
Tags
On-Demand
