Metadata-Version: 2.1
Name: ptcn
Version: 0.1.0
Summary: TF2 implementation of a Temporal Convolutional Network with a probabilistic twist
Home-page: https://github.com/UpstatePedro/ptcn
License: GPL-3.0-only
Author: UpstatePedro
Requires-Python: >=3.8,<4.0
Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3)
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.8
Requires-Dist: tensorflow (>=2.2.0,<3.0.0)
Requires-Dist: tensorflow-addons (>=0.10.0,<0.11.0)
Project-URL: Repository, https://github.com/UpstatePedro/ptcn
Description-Content-Type: text/markdown

# ptcn

Tensorflow (2.x) implementation of a Temporal Convolutional Network architecture, with a probabilistic twist.

This project indulges a couple of curiosities:

1. Working with convolutional sequence-to-sequence models a la [An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling](https://arxiv.org/abs/1803.01271)
2. Adding a bayesian twist to the network a la [Bayesian Segnet: Model Uncertainty in Deep Convolutional Encoder-Decoder Architectures for Scene Understanding](https://arxiv.org/abs/1511.02680)

This implementation has been inspired by other projects, including:
- https://github.com/locuslab/TCN
- https://github.com/Baichenjia/Tensorflow-TCN
- https://github.com/philipperemy/keras-tcn

