torchex 0.0.10

[](https://github.com/0h-n0/torchex)
[](https://github.com/0h-n0/torchex)
[](https://badge.fury.io/py/torchex)
[](https://circleci.com/gh/0h-n0/torchex)
[](https://travis-ci.org/0h-n0/torchex)
[](https://codecov.io/gh/0h-n0/torchex)
[](https://torchex.readthedocs.io/en/latest/?badge=latest)
[](https://codeclimate.com/github/0h-n0/torchex/maintainability)
[](https://codeclimate.com/github/0h-n0/torchex/test_coverage)
[](https://bettercodehub.com/)
[](https://pypi.org/project/torchex/)

# (WIP) `torchexq library

`torchex` library provides advanced Neural Network Layers. You can easily use them like using original pytorch.

## Installation

```
$ pip install torchex
```

## Requirements

* Pytorch >= 0.4.1

## Documentation

* https://torchex.readthedocs.io/en/latest/index.html

## How to use

### Lazy Modules

```python
import torch
import torchex.nn as exnn

net = exnn.Linear(10)
# You don't need to give the size of input for this module.
# This network is equivalent to `nn.Linear(100, 10)`.

x = troch.randn(10, 100)

y = net(x)
```


## TODO

### Layers

- [x] support fundamental complex operations
- to_complex method
- to_real method
- complex_norm method
- [ ] add submodule for many examples.
- [ ] SeparableConv2D
- [ ] LocallyConnected1D
- [x] Highway
- [x] Inception
- [x] InceptionBN
- [x] Conv2dLocal
- [x] MLPConv2d
* [Network In Network](https://arxiv.org/abs/1312.4400v3)
- [ ] NaryTreeLSTM
- [ ] StatefulZoneoutLSTM
- [ ] StatefulPeepholeLSTM
- [ ] StatefulMGU
- [ ] BinaryHierarchicalSoftmax
- [ ] BlackOut
- [ ] CRF1d
- [ ] SimplifiedDropconnect
- [ ] Swish
- [ ] NegativeSampling
- [ ] ResidualCell
- [ ] Attention Cell
* [XiaoIce Band: A Melody and Arrangement Generation Framework for Pop Music](https://www.kdd.org/kdd2018/accepted-papers/view/xiaoice-banda-melody-and-arrangement-generation-framework-for-pop-music)
- [ ] MLP Cell
* same as above.
- [ ] DFT2d
* [Rotation Equivariance and Invariance in Convolutional Neural Networks](https://arxiv.org/pdf/1805.12301.pdf)
* https://github.com/bchidest/RiCNN/tree/master/ricnn
- [ ] My original DFT layer (made by Koji Ono)
- [ ] DFT1d
- [x] DFT2d
- [ ] DFT3d
- [ ] iDFT1d
- [ ] iDFT2d
- [ ] iDFT3d
- [ ] RFFT1d
- [ ] RFFT2d
- [ ] RFFT3d

- [ ] Conic Convolutional Layers
* same as above.

### Zoo

- [x] ImageTransferNet

### Optimizer

- [ ] chainer.optimizer_hooks.GradientLARS

### Atiributions

- [x] Integrated Gradients

## Examples


## Related Projects

* torchhp
* Hyper-Parameter Turning Library for Pytorch.
* torchrl
* Pytorch Reinforcement Learning Library.
* torchchem
* TorchChem aims to provide a high quality open-source toolchain that democratizes the use of deep-learning in drug discovery, materials science, quantum chemistry, and biology.
* torchml
* Auto model optimization library for pytorch.
* torcdata
* Pytorch Datasets.

## Codes References

* Chainer
* One of the most wonderfull DeepLearning framework.
* https://github.com/chainer/chainer
* NLP
* allenNLP
* https://github.com/allenai/allennlp
* fairseq
* https://github.com/pytorch/fairseq
* text
* https://github.com/pytorch/text
* translate
* https://github.com/pytorch/translate
* Audio
* neural_sp
* https://github.com/hirofumi0810/neural_sp/
* deepspeech.pytorch
* https://github.com/SeanNaren/deepspeech.pytorch
* Awesome Speech Recognition Speech Synthesis Papers
* https://github.com/zzw922cn/awesome-speech-recognition-speech-synthesis-papers
* speech
* https://github.com/awni/speech
* pytorch-asr
* https://github.com/jinserk/pytorch-asr