deepware-ai / Q-MANN

Quantized Memory-Augmented Neural Networks (AAAI-18)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Q-MANN

Quantized Memory-Augmented Neural Networks (AAAI-18) (https://www.aaai.org/ocs/index.php/AAAI/AAAI18/paper/download/16732/16638)

Description

Memory networks based model

Fixed-point quantization

Training & inference w/ quatized model

C & Cuda based deep learning libaray

Attention model

Hardware friendly attnetion model

Dot product attention -> hamming distance attnetion

How to make & run

run.sh (in ./MemN2N)

Configurations

define.h (in ./MemN2N)

bAbI dataset

to be updated

Citation

Please cite our paper in your publications if it helps your research work:

@inproceedings{park2018quantized,
  title={Quantized memory-augmented neural networks},
  author={Park, Seongsik and Kim, Seijoon and Lee, Seil and Bae, Ho and Yoon, Sungroh},
  booktitle={Thirty-Second AAAI Conference on Artificial Intelligence},
  year={2018}
}

About

Quantized Memory-Augmented Neural Networks (AAAI-18)

License:MIT License


Languages

Language:C 73.7%Language:Cuda 23.9%Language:Python 1.1%Language:Shell 0.8%Language:Makefile 0.5%Language:C++ 0.1%