notorch 0.2.3a0

Neural networks and ML from scratch in pure Python

NoTorch 🕯

A pure Python, educational machine learning library — built from scratch.

NoTorch is a learning-focused machine learning and deep learning library that re-implements popular models using only core Python (and pandas for data handling). Inspired by the design of scikit-learn and PyTorch, NoTorch offers a simple and consistent API while prioritizing readability, clarity, and educational value over performance or production use.

This project is being developed as part of a personal journey to deeply understand how machine learning works — and to help others learn by example. Whether you're a student, a hobbyist, or just curious about what happens under the hood of your favorite ML libraries, NoTorch is here to make complex concepts transparent and approachable.

---

🔍 Project Goals

• ✅ Build core ML and DL models from scratch in pure Python
• ✅ Keep the code readable, modular, and well-documented
• ✅ Mimic the familiar APIs of sklearn and torch for a smoother learning curve
• ✅ Prioritize educational clarity over performance or abstraction
• ✅ Share the learning journey with the community

---

✅ Implemented Models

Classical Machine Learning

• K-Nearest Neighbors (KNN)
• Decision Tree
• Random Forest
• Naive Bayes
• Linear Regression
• Logistic Regression
• Support Vector Machine (SVM)

Neural Networks (Coming Soon)

• Basic Feedforward Neural Network
• Manual Backpropagation Engine
• Custom Autograd System (Experimental)
• Convolutional Neural Networks (CNNs)
• Recurrent Neural Networks (RNNs)
• Long Short-Term Memory (LSTM)

---

🛠 Example Usage

KNN Model

from NoTorch.KNN import KNN

model = KNN(n_neighbors=1)
model.fit(X_train, y_train)

print(model.predict(X_test))

---

🚧 This Project is Under Active Development

NoTorch is a work-in-progress, built step-by-step with the intent to learn by building. Contributions, ideas, and educational feedback are very welcome!