PyTorch Deep Learning Journey 🚀

A comprehensive exploration of computer vision using PyTorch

This repository documents my self-directed learning journey into deep learning and computer vision, featuring hands-on implementations of convolutional neural networks and comprehensive PyTorch fundamentals. Built as part of my preparation for university computer science studies.

🎯 Repository Overview

This project showcases practical machine learning skills through two main components:

Core Projects - Real-world CNN implementations

• EMNIST Digit Recognition: Custom CNN for handwritten digit classification
• CIFAR-10 Object Classification: Advanced CNN with modern techniques

Learning Modules - Progressive skill building

• Fundamentals: Tensor operations, neural network basics
• Advanced Concepts: Training loops, optimization, model evaluation

---

Project Structure

pytorch-learning-journey/
├── 🎯 projects/
│   ├── emnist-digit-classification/     # Handwritten digit recognition
│   │   ├── EMNIST_Dataset.py
│   │   └── README.md
│   └── cifar10-image-classification/    # Natural image classification
│       ├── CIFAR-10_Dataset.py
│       └── README.md
├── 📚 learning-exercises/               # Progressive learning modules
│   ├── 00_pytorch_basics.py
│   ├── 01_binary_classification.py
│   ├── 02_binary_classification_exercise.py
│   ├── 03_multilcass_classification.py
│   ├── 04_multilcass_classification_exercise.py
│   ├── 05_computervision.py
│   └── 06_custom_datasets.py
├── 📊 assets/                          # Visualizations and diagrams
│   └── images/
└── 📖 README.md

---

📈 Results & Achievements

Model Performance

• Successfully implemented two distinct CNN architectures
• Achieved competitive accuracy on standard benchmarks
• Demonstrated understanding of regularization and optimization

Training Insights

• Effective loss curves showing proper convergence
• No overfitting through proper regularization
• Efficient training loops with progress tracking

Technical Skills Developed

Category	Skills Acquired
Deep Learning	CNN architecture design, training loops, optimization
PyTorch	Model definition, data loading, GPU acceleration
Computer Vision	Image preprocessing, data augmentation, evaluation
Software Engineering	Clean code, documentation, version control

---

📚 Learning Resources

Primary References:

• MIT 6.S191: Introduction to Deep Learning - MIT's deep learning course
• PyTorch Course by Daniel Bourke - 24hours PyTorch course
• PyTorch Official Tutorials - Comprehensive framework guide

Datasets:

• EMNIST - Extended MNIST handwritten characters
• CIFAR-10 - Natural image classification benchmark

---

🤝 Connect & Collaborate

I'm passionate about machine learning and always excited to discuss projects, share knowledge, or collaborate on interesting problems!