https://github.com/onlytailei/Value-Iteration-Networks-PyTorch

PyTorch implementation of the Value Iteration Networks (VIN) (NIPS '16 best paper)
https://github.com/onlytailei/Value-Iteration-Networks-PyTorch

Last synced: about 2 months ago
JSON representation

PyTorch implementation of the Value Iteration Networks (VIN) (NIPS '16 best paper)

• Host: GitHub
• URL: https://github.com/onlytailei/Value-Iteration-Networks-PyTorch
• Owner: onlytailei
• Created: 2017-03-13T10:04:39.000Z (over 7 years ago)
• Default Branch: master
• Last Pushed: 2017-03-13T13:46:31.000Z (over 7 years ago)
• Last Synced: 2024-05-22T07:52:35.272Z (4 months ago)
• Language: Python
• Homepage:
• Size: 431 KB
• Stars: 78
• Watchers: 5
• Forks: 23
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# [Value Iteration Networks](https://arxiv.org/abs/1602.02867) in PyTorch

> Tamar, A., Wu, Y., Thomas, G., Levine, S., and Abbeel, P. _Value Iteration Networks_. Neural Information Processing Systems (NIPS) 2016

This repository contains an implementation of Value Iteration Networks (VIN) in PyTorch based on the original [Theano implementation](https://github.com/avivt/VIN) by the authors and the [TensoFlow implementation](https://github.com/TheAbhiKumar/tensorflow-value-iteration-networks) by [Abhishek Kumar](https://github.com/TheAbhiKumar/tensorflow-value-iteration-networks).

VIN won the Best Paper Award at NIPS 2016. 

![Value Iteration Network and Module](img/vin.png)

## Dependencies

* Python 2.7

* PyTorch 

* SciPy >= 0.18.1 (to load the data)

## Datasets

- The datasets is from the [author's repository](https://github.com/avivt/VIN/tree/master/data). This repository contains the 8x8 GridWorld dataset for convenience and its small size.

- utils.py and data.py are from [Abhishek Kumar's repository](https://github.com/TheAbhiKumar/tensorflow-value-iteration-networks)

## Training

```

python train_main.py

```

Several arguments can be set in train_main.py like learning rate. Please check train_main.py for details.

```

python train_main.py --lr 0.001

```

## References

* [Value Iteration Networks on arXiv](https://arxiv.org/abs/1602.02867)

* [Aviv Tamar's (author) implementation in Theano](https://github.com/avivt/VIN)

* [Abhishek Kumar's implementation in TensorFlow](https://github.com/TheAbhiKumar/tensorflow-value-iteration-networks).