MUNet

MUNet: Macro Unit-based Convolutional Neural Networks for Mobile Devices

Dependencies

• Python 3.5+
• Tensorflow 1.2.0+ (tested upto version 1.14)
• Keras 2.0.6+ (tested upto version 2.2.4)

Usage

1. Cloning the repository

$ git clone https://github.com/kdhht2334/MUNet-CVPR18-Workshop.git
$ cd MUNET-CVPR18-Workshop/

2. Run

$ python main.py --gpus gpu-id

For example, gpu-id is from 0 to 3 in case of 4GPUs device.

If you use 1GPU device, then gpu-id is 0.

Milestone

• Cifar dataset experiments
• SVHN dataset experiments
• Tiny ImageNet dataset experiments
• Add results table

Citation

If this work is helpful for your research, please cite our:

 @InProceedings{Kim_2018_CVPR_Workshops,
author = {Ha Kim, Dae and Hyun Lee, Seung and Cheol Song, Byung},
title = {MUNet: Macro Unit-Based Convolutional Neural Network for Mobile Devices},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
month = {June},
year = {2018}
}