micrograd is a minimalistic neural network library designed to facilitate an understanding of the inner workings of neural networks. This library provides a simple interface for designing neural network architectures, executing a forward pass, and performing backpropagation. It includes a small set of neurons organized into a network, and utilizes a directed acyclic graph for visualizing state propagation. This visualization aids in the step-by-step, layer-by-layer understanding of loss propagation during the backward pass.
- Design and implement simple neural network architectures.
- Perform forward passes and backpropagation.
- Visualize network state and loss propagation using directed acyclic graphs.
Before installing micrograd
, ensure you have the following prerequisites installed:
- Python 3.6 or higher
- Relevant Python libraries:
numpy
,matplotlib
for visualization (if necessary)
-
Clone the repository to your local machine:
git clone https://github.com/[your-username]/micrograd.git
-
Navigate to the working directory:
cd micrograd
-
Install the required dependencies:
pip install -r requirements.txt
Contributions make the open-source community a fantastic place to learn, inspire, and create. Any contributions you make to this project are deeply appreciated.
- Fork the Project
- Create your Feature Branch (
git checkout -b feature/AmazingFeature
) - Commit your Changes (
git commit -m 'Add some AmazingFeature'
) - Push to the Branch (
git push origin feature/AmazingFeature
) - Open a Pull Request
This project is licensed under the MIT License - see the LICENSE
file for details.
- Name - Jay Joshi
- Email - [email protected]
- Project Link: https://github.com/jyjoshi/micrograd
- Andrej Karpathy: https://github.com/karpathy/