gpkg.keras.mnist 0.4.0

MNIST related models in Keras

Models
######

mnist-acgan
###########

*Auxiliary Classifier Generative Adversarial Network (ACGAN) for MNIST in
Keras*

Operations
==========

train
^^^^^

*Train the ACGAN*

Flags
-----

**batch-size**
*Training batch size (default is 100)*

**beta-1**
*Beta 1 (default is 0.5)*

**epochs**
*Number of epochs to train (default is 100)*

**learning-rate**
*Learning rate (default is 0.0002)*

References
==========

- https://github.com/keras-team/keras/blob/master/examples/mnist_acgan.py
- https://arxiv.org/abs/1511.06434

mnist-cnn
#########

*Convolutional neural network (CNN) classifier for MNIST in Keras*

Operations
==========

train
^^^^^

*Train the CNN*

Flags
-----

**batch-size**
*Training batch size (default is 128)*

**epochs**
*Number of epochs to train (default is 12)*

References
==========

- https://github.com/keras-team/keras/blob/master/examples/mnist_cnn.py

mnist-denoising-autoencoder
###########################

*Denoising autoencoder for MNIST in Keras*

Operations
==========

train
^^^^^

*Train the autoencoder*

Flags
-----

**batch-size**
*Training batch size (default is 128)*

**epochs**
*Number of epochs to train (default is 30)*

References
==========

- https://github.com/keras-team/keras/blob/master/examples/mnist_denoising_autoencoder.py

mnist-hierarchical-rnn
######################

*Hierarchical RNN (HRNN) classifier for MNIST in Keras*

Operations
==========

train
^^^^^

*Train the HRNN*

Flags
-----

**batch-size**
*Training batch size (default is 32)*

**epochs**
*Number of epochs to train (default is 5)*

References
==========

- https://github.com/keras-team/keras/blob/master/examples/mnist_hierarchical_rnn.py
- https://arxiv.org/abs/1506.01057
- http://ieeexplore.ieee.org/document/7298714/

mnist-irnn
##########

*Implementation of 'A Simple Way to Initialize Recurrent Networks of Rectified
Linear Units' with MNIST in Keras*

Operations
==========

train
^^^^^

*Train the RNN*

Flags
-----

**batch-size**
*Training batch size (default is 32)*

**epochs**
*Number of epochs to train (default is 200)*

**learning-rate**
*Learning rate (default is 1e-06)*

References
==========

- https://github.com/keras-team/keras/blob/master/examples/mnist_irnn.py
- http://arxiv.org/pdf/1504.00941v2.pdf

mnist-mlp
#########

*Multilayer perceptron (MLP) classifier for MNIST in Keras*

Operations
==========

train
^^^^^

*Train the MLP*

Flags
-----

**batch-size**
*Training batch size (default is 128)*

**epochs**
*Number of epochs to train (default is 20)*

References
==========

- https://github.com/keras-team/keras/blob/master/examples/mnist_mlp.py

mnist-net2net
#############

*Implementation of 'Net2Net: Accelerating Learning via Knowledge Transfer'
with MNIST in Keras*

Operations
==========

train
^^^^^

*Train the network*

Flags
-----

**batch-size**
*Training batch size (default is 32)*

**epochs**
*Number of epochs to train (default is 3)*

References
==========

- https://github.com/keras-team/keras/blob/master/examples/mnist_net2net.py
- http://arxiv.org/abs/1511.05641

mnist-siamese
#############

*Siamese MLP classifier for MNIST in Keras*

Operations
==========

train
^^^^^

*Train the MLP*

Flags
-----

**batch-size**
*Training batch size (default is 128)*

**epochs**
*Number of epochs to train (default is 20)*

References
==========

- https://github.com/keras-team/keras/blob/master/examples/mnist_siamese.py
- http://yann.lecun.com/exdb/publis/pdf/hadsell-chopra-lecun-06.pdf

mnist-swwae
###########

*Stacked what-where autoencoder for MNIST in Keras*

Operations
==========

train
^^^^^

*Train the MLP*

Flags
-----

**batch-size**
*Training batch size (default is 128)*

**epochs**
*Number of epochs to train (default is 5)*

**pool-size**
*kernel size used for the MaxPooling2D (default is 2)

Choices:
2
3

*

References
==========

- https://github.com/keras-team/keras/blob/master/examples/mnist_swwae.py
- https://arxiv.org/abs/1311.2901v3
- https://arxiv.org/abs/1506.02351v8