Binary Neural Networks on PyTorch

This repository implements three popular papers that introduced the concept of Binary Neural Networks:

• XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networks.
• Binarized Neural Networks
• DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients

The project is organized as follows:

• models folder contains CNN models (simple mlp, Network-in-Network, LeNet5, etc.)
• classifiers/{type}_classifier.py contains the test and train procedures; where type = {bnn, xnor, dorefa}
• models/{type}_layers.py contains the binarylayers implementation (binary activation, binary conv and fully-connected layers, gradient update); where type = {bnn, xnor, dorefa}
• yml folder contains configuration files with hyperparameters
• main.py represents the entry file

Installation

All packages are in requirement.txt Install the dependencies:

pip install -r requirements.txt

Basic usage

$ python main.py app:{yml_file}

Example

Network-in-Network on CIFAR10 dataset. All hyper parameters are in .yml file.

$ python main.py app:yml/nin_cifar10.yml

Related Applications

If you find this code useful in your research, please consider citing one of the works in this section.

• Fast and Accurate Inference on Microcontrollers With Boosted Cooperative Convolutional Neural Networks (BC-Net)
• CoopNet: Cooperative Convolutional Neural Network for Low-Power MCUs
• TentacleNet: A Pseudo-Ensemble Template for Accurate Binary Convolutional Neural Networks

License

MIT