Metadata-Version: 2.1
Name: pytcn
Version: 0.0.2
Summary: Temporal Convolution Network (TCN) based on Keras
Home-page: https://github.com/zhf026
Author: zhf026
Author-email: zhf026@outlook.com
License: UNKNOWN
Platform: UNKNOWN
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.6
Description-Content-Type: text/markdown
Requires-Dist: keras
Requires-Dist: tensorflow

# Temporal Convolution Network (TCN)

## Description
This is a implement of temporal Convolution Network (TCN) by using keras, and the 
version uses a dense layer as the output layer instead of fully convolution network (FCN) structure depicted in paper "An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling".


## Usage
```python
    import keras 
    from tcn import *

    # Build TCN model
    # model is a keras model
    model = tcn(input_shape=(24, 7), n_blocks=3, filters=[16, 32, 64], 
                kernel_size=[24, 24, 24], dropout_rate=[0.1, 0.1, 0.1], 
                n_outputs=6)
    # Plot the structure of model
    keras.utils.plot_model(model, to_file='model.png', show_shapes=True)
```

![](./assets/model.png)

## References
> [An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling](https://arxiv.org/pdf/1803.01271.pdf)



