MatthieuCourbariaux/BinaryNet

Stars
1,040
Rank 44,297 (Top 0.9 %)
Language
Python
License
BSD 3-Clause "New...
Created almost 9 years ago
Updated almost 6 years ago

MatthieuCourbariaux/BinaryNet

MatthieuCourbariaux

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Training Deep Neural Networks with Weights and Activations Constrained to +1 or -1

BinaryNet

This repository enables the reproduction of the experiments described in the article:
BinaryNet: Training Deep Neural Networks with Weights and Activations Constrained to +1 or -1.

It is divided in two subrepositories:

Train-time enables the reproduction of the benchmark results reported in the article
Run-time demonstrates the XNOR and baseline GPU kernels described in the article

BinaryConnect

Training Deep Neural Networks with binary weights during propagations

deep-learning-multipliers

Training deep neural networks with low precision multiplications

8-bit-deep-learning

Training neural networks with 8-bit computations

AdversarialGradient

A very simple variant of adversarial training that yields excellent results on MNIST

binary-matrix-product

Fast binary matrix product on CPU

Eigen-MLP

Training a Maxout MLP on MNIST with a CPU

ift6266

ift6266 project