TensorWeaver

A modern educational deep learning framework for students, engineers and researchers

Features • Installation • Examples • Documentation • License • Acknowledgments

Introduction

TensorWeaver is a deep learning framework designed specifically for students, engineers and researchers who want to understand how deep learning frameworks work under the hood. Unlike industrial frameworks like PyTorch and TensorFlow that prioritize performance and scalability, TensorWeaver focuses on clarity, readability, and simplicity.

Built entirely in Python with only NumPy as a dependency, TensorWeaver's codebase is transparent and approachable, making it an ideal learning resource for those who want to demystify the "magic" behind modern AI frameworks.

The target users are students, engineers and researchers who want to fully understand the working principles of deep learning frameworks and gain the skills to debug, extend and optimize their own projects with any deep learning framework.

Features

• Purely Educational: Designed from the ground up as a learning tool with clear, well-documented code
• PyTorch-like API: Familiar interface reduces learning curve and eases transition to industrial frameworks
• Lightweight and Readable: Built with pure Python and minimal dependencies (merely NumPy)
• Fully Functional: Supports essential deep learning components:
  • Automatic differentiation engine
  • Common neural network operators
  • Loss functions and optimizers
  • Model definition and training
• Advanced Capabilities:
  • ONNX export functionality
• Comprehensive Documentation: Detailed explanations of implementation details and design choices

Installation

# Install from PyPI
pip install tensorweaver

# Or install from source
git clone https://github.com/howl-anderson/tensorweaver.git
cd tensorweaver
poetry install

see poetry for more details if you don't have poetry installed.

Examples

See milestones for examples.

Documentation

see https://www.tensorweaver.ai

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• The project draws inspiration from educational frameworks like Micrograd, TinyFlow, and DeZero
• Special thanks to the open-source deep learning community for their pioneering work
• Thanks to all contributors and users who help improve this educational resource