VQ-VAE Keras Implementation

Keras implementaion of VQ-VAE (Vector Quantizer Variational AutoEncoder)

Sources:

• https://arxiv.org/abs/1711.00937
• https://github.com/deepmind/sonnet/blob/master/sonnet/python/modules/nets/vqvae.py
• https://github.com/deepmind/sonnet/blob/master/sonnet/examples/vqvae_example.ipynb

Example Notebook:

Original vs Reconstructed

Notes on compatibility

• The notebook was created on a Google Colab machine (GPU accelerated) which ran TensorFlow version 1.x

• The notebook was tested with TensorFlow version 2.2.0 and Keras version 2.3.1 on a Google Colab machine (GPU accelerated) and worked when removing the parameter validate_indices from the call tf.nn.embedding_lookup (it is a deprecated argument).

• If you have issues with eager execution with TensorFlow version 2.x or higher, issue #2 might help you.