NeuralMerger

Official implementation of Unifying and Merging Well-trained Deep Neural Networks for Inference Stage.

Created by Yi-Min Chou , Yi-Ming Chan, Jia-Hong Lee, Chih-Yi Chiu, Chu-Song Chen

Usage

Fine-tuning: Finetune the merged model of two well-trained neural networks (Tensorflow implementation).

Inference: Test the speed of the merged model (C implementation).

NeuralMerger
    ├─────── Fine-tuning
    └─────── Inference

1.Clone the NeuralMerger repository:

$ git clone --recursive https://github.com/ivclab/NeuralMerger.git

2.Follow the instruction in Fine-tuning and get the well-trained merged model.

3.Test the well-trained merged model on Inference.

Citation

Please cite following paper if these codes help your research:

@inproceedings{chou2018unifying,
  title={Unifying and merging well-trained deep neural networks for inference stage},
  author={Chou, Yi-Min and Chan, Yi-Ming and Lee, Jia-Hong and Chiu, Chih-Yi and Chen, Chu-Song},
  booktitle={Proceedings of the 27th International Joint Conference on Artificial Intelligence},
  pages={2049--2056},
  year={2018},
  organization={AAAI Press}
}

@inproceedings{chou2018merging,
  title={Merging Deep Neural Networks for Mobile Devices},
  author={Chou, Yi-Min and Chan, Yi-Ming and Lee, Jia-Hong and Chiu, Chih-Yi and Chen, Chu-Song},
  booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops},
  pages={1686--1694},
  year={2018}
}   

Contact

Please feel free to leave suggestions or comments to Yi-Min Chou([email protected]) , Yi-Ming Chan([email protected]), Jia-Hong Lee([email protected]), Chih-Yi Chiu([email protected]), Chu-Song Chen([email protected])