🚀 AI & Machine Learning From Scratch Series

Welcome to the AI & Machine Learning From Scratch series! This repository is dedicated to breaking down and building foundational Artificial Intelligence and Machine Learning models, algorithms, and neural networks from first principles.

Instead of treating modern frameworks like black boxes, this series explores how the math, algorithms, optimization routines, and backpropagation actually work under the hood using pure Python, NumPy, and PyTorch fundamentals.

---

📚 Repository Roadmap & Topics

#	Topic	Notebook	Colab	Description	Tech Stack	Status
1	Linear Regression	NeuralNetwork/LinearRegression/		Single-variable linear regression predicting marks from study hours	PyTorch (nn.Module), Pandas, Matplotlib	✅ Completed
2	Neural Network from Scratch	NeuralNetwork/numpy_mnist.ipynb		Multi-class digit classification on the MNIST dataset	Pure NumPy, Math & Linear Algebra	✅ Completed
3	Logistic Regression	Coming Soon	—	Binary and multiclass classification with sigmoid & cross-entropy loss	NumPy / PyTorch	⏳ In Progress
4	Multi-Layer Perceptron (MLP)	Coming Soon	—	Fully-connected deep neural networks with activation functions & backpropagation	NumPy / PyTorch	📅 Planned
5	Optimizers From Scratch	Coming Soon	—	Implementing Gradient Descent, Momentum, RMSprop, and Adam	Pure Python & NumPy	📅 Planned
6	Convolutional Neural Networks (CNN)	Coming Soon	—	Convolutions, pooling, feature maps, and computer vision classification	PyTorch	📅 Planned
7	Transformers & Attention	Coming Soon	—	Self-attention mechanism, Multi-head attention, and Encoder-Decoder architecture	PyTorch	📅 Planned

---

🛠️ Getting Started

1. Prerequisites

Make sure you have Python 3.10+ installed on your system.

2. Clone the Repository

git clone https://github.com/prachin77/From-Scratch.git
cd From-Scratch

3. Set Up a Virtual Environment

# Windows
python -m venv venv
venv\Scripts\activate

# macOS / Linux
python3 -m venv venv
source venv/bin/activate

4. Install Dependencies

pip install torch numpy pandas matplotlib scikit-learn jupyter

5. Launch Jupyter Lab / Notebook

jupyter notebook

Navigate to any notebook (e.g. NeuralNetwork/LinearRegression/linear_nn.ipynb) and run the cells interactively!

---

🤝 How to Engage & Contribute

Community contributions and discussions are welcome! Whether you are catching a typo, implementing a new algorithm, optimizing a function, or suggesting a topic:

💬 GitHub Discussions

• Have a question about how backpropagation or gradient descent works?
• Want to suggest the next topic or model to implement?
• Head over to the Discussions tab to share ideas, ask questions, and learn together.

🐛 Reporting Issues

• If you find a bug, mathematical inaccuracy, or broken link:
  1. Go to the Issues tab.
  2. Click New Issue.
  3. Provide a clear description of the problem, code snippet, and steps to reproduce.

🔀 Contributing via Pull Requests (PRs)

1. Fork this repository to your GitHub account.
2. Clone your fork locally:

   git clone https://github.com/prachin77/From-Scratch.git

3. Create a new branch for your feature or fix:

   git checkout -b feature/topic-name-or-fix

4. Make your changes:
   • Keep notebooks clean and well-commented.
   • Include mathematical intuition and markdown explanations where appropriate.
   • Ensure any heavy datasets or raw files are added to .gitignore.
5. Commit your changes:

   git add .
   git commit -m "Add: [Topic/Algorithm] implementation from scratch"

6. Push to your fork:

   git push origin feature/topic-name-or-fix

7. Open a Pull Request (PR) against the main branch with a description of what you added or improved.

---

📄 License

This repository is open-source and available under the MIT License.

---

⭐ If you find this series helpful, consider giving it a star!