This repository contains the coursework for Carnegie Mellon's Intro to Deep Learning (11-685) class.
Throughout the course, I implemented a wide range of deep learning models from scratch and in PyTorch, gradually building up from foundational concepts to modern architectures used in real-world AI systems.
The class began with the fundamentals — including multi-layer perceptrons (MLPs) — and moved into more advanced topics like convolutional neural networks (CNNs), recurrent neural networks (RNNs), attention mechanisms, and transformers.
Each homework had two parts:
- Part 1 focused on building neural network architectures from the ground up using pure Python and NumPy.
- Part 2 involved using PyTorch to train, fine-tune, and evaluate models on real tasks.
Assignments included tasks like sequence classification, image recognition, speech-to-text, and language modeling — providing hands-on experience with both architecture design and model training.
Austin Windham
GitHub Profile