A lightweight tensor computation library with automatic differentiation, implemented from scratch in pure Python. Perfect for educational purposes and understanding the fundamentals of deep learning frameworks.
🧠 Key Features:
- Eager evaluation model
- Minimal autograd engine
- Multi-Layer Perceptron (MLP) support
- Lightweight
- Educational-focused design
mini-tensor/
├── tensor.py # Core tensor class with autograd
├── mlp.py # Neural network implementation
├── util.py # Activation functions & optimizers
└── requirements.txt