Welcome to my personal learning laboratory for Machine Learning and Deep Learning. This repository documents my step-by-step progress, starting from raw Python logic up to training Neural Networks using PyTorch.
Here is the chronological order of my learning journey and the scripts included in this repository:
- Tensor Operations: Exploring PyTorch tensors, matrix multiplications, and shapes.
- Linear Regression: Implementing basic predictive models from scratch.
- 🎯 MNIST Digit Classifier: A Multi-Layer Perceptron (MLP) / Convolutional Neural Network (CNN) that recognizes handwritten digits (0-9) with high accuracy.
- 👕 Fashion-MNIST Classifier: A deeper model trained to classify articles of clothing (T-shirts, trousers, sneakers, etc.) into 10 distinct categories.
- Framework: PyTorch (Tensors, Autograd,
nn.Module, DataLoader) - Optimization: Stochastic Gradient Descent (SGD), Adam Optimizer, and Cross-Entropy Loss.
- Hardware: Tracking performance and moving tensors between CPU and GPU (
cuda).
I built this repository to keep track of my deep learning journey. My ultimate goal is to understand how modern AI models process data behind the scenes, master optimization techniques, and move toward building custom production-ready computer vision applications.
Feel free to look around the code! If you have any suggestions on improving training loops or model architectures, please open an issue! ⭐