ARAE

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #ai #typescript #llm #go #golang #android #machine-learning #rust #deep-learning #linux

★ 18 Open GitHub ↗

tensorflow implementation of Adversarially Regularized Autoencoders (ARAE)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# ARAE
tensorflow implementation of [Adversarially Regularized Autoencoders for Generating Discrete Structures](https://arxiv.org/abs/1706.04223) (ARAE)

While the Paper used the Stanford Natural Language Inference dataset for the text generation, this implementation only used the mnist dataset.
I implemented the continuous version for this implementaion, but discrete version is implemented as footnote in this code.

## Dependencies

1. tensorflow == 1.0.0
1. numpy == 1.12.0
1. matplotlib == 1.3.1

## Steps

Run the following code for image reconstruction.

<pre><code>
python train.py
</code></pre>

## Results

- The model trained 100000 steps

- Generated from fake(noise) data

 <img src="https://raw.githubusercontent.com/soobin3230/ARAE/master/png/fake_76500.png" width="112"/>
 <img src="https://raw.githubusercontent.com/soobin3230/ARAE/master/png/fake_91500.png" width="112"/>
 <img src="https://raw.githubusercontent.com/soobin3230/ARAE/master/png/fake_99500.png" width="112"/>

The result from noise tend to appear several figures simultaneously.

- Generated from real data

 <img src="https://raw.githubusercontent.com/soobin3230/ARAE/master/png/real_42000.png" width="112"/>
 <img src="https://raw.githubusercontent.com/soobin3230/ARAE/master/png/real_42000.png" width="112"/>
 <img src="https://raw.githubusercontent.com/soobin3230/ARAE/master/png/real_42000.png" width="112"/>

## Notes

I didn't multiply the critic gradient before backpropping to the encoder.

ARAE

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README