High-Capacity Expert Binary Networks (ICLR 2021)

This code provides the core components for building networks based on the architectures and Expert Binary Convolutional Block introduced in the High-Capacity Expert Binary Networks paper. You can find the full version of the paper here.

1. Installation

To install and test the code simply clone the current repo, install the required packages listed bellow and prepare the training and/or testing(validation) data.

1.1 Requirements

torch >= 1.6.0 torchvision >= 0.5.0 bnn >= 0.1.1

1.2. Data preparation

Download the ImageNet dataset from the official webpage, creating a folder with the following structure:

│imagenet/
├──train/
│  ├── n01440764
│  │   ├── n01440764_10026.JPEG
│  │   ├── n01440764_10027.JPEG
│  │   ├── ......
│  ├── ......
├──val/
│  ├── n01440764
│  │   ├── ILSVRC2012_val_00000293.JPEG
│  │   ├── ILSVRC2012_val_00002138.JPEG
│  │   ├── ......
│  ├── ......

You can construct this structure using for example the script found here.

2. Testing pretrained model

A packed model, with the extra training components removed (i.e. the gate selection is performed using selection instead of multiplication as its the case for training etc) is available for download here.

python test/test.py imagenet_valid_location --path-to-model download_model_path

Reference

If you find this repo useful, please consider citing:

@inproceedings{bulat2021high,
  title={High-Capacity Expert Binary Networks},
  author={Bulat, Adrian and Martinez, Brais and Tzimiropoulos, Georgios},
  booktitle={International Conference on Learning Representations (ICLR)},
  year={2021}
}