https://github.com/seujung/relational-network-gluon

Gluon implement of "A simple neural network module of relational reasoning"
https://github.com/seujung/relational-network-gluon

deep-learning gluon mxnet relational-networks vqa

Last synced: 3 days ago
JSON representation

Gluon implement of "A simple neural network module of relational reasoning"

• Host: GitHub
• URL: https://github.com/seujung/relational-network-gluon
• Owner: seujung
• Created: 2018-07-30T14:03:08.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2018-08-20T00:49:19.000Z (over 6 years ago)
• Last Synced: 2024-11-08T18:45:57.432Z (5 months ago)
• Topics: deep-learning, gluon, mxnet, relational-networks, vqa
• Language: Jupyter Notebook
• Homepage:
• Size: 3.28 MB
• Stars: 9
• Watchers: 2
• Forks: 2
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• Awesome-MXNet - relational-network-gluon

README

        
# relational network with Gluon

---

Gluon inplementation of [A simple neural module for relational reasoning](https://arxiv.org/abs/1706.01427)

![net_structure](assets/network_structure.png)

## Requirements

- Python 3.6.1

- Mxnet(1.2)

- tqdm

- opencv-python

## Usage

- arguments

  - batch_size : Define batch size (defualt=64)

  - epoches : Define total epoches (default=50)

  - GPU_COUNT : Use GPU count (default=2)

  - show_status : show loss and accuracy for each epoch (default=True)

###### data generation

```

python sort-of-clevr.py

``` 

###### default setting

```

python main.py

``` 

or

###### manual setting

```

python main.py --batch_size=32 --epoches=100

```

## Results

![perf_acc](assets/perf_result_acc.png)

![perf_loss](assets/perf_result_l.png)

## Reference

- https://github.com/kimhc6028/relational-networks