Class List

Here are the classes, structs, unions and interfaces with brief descriptions:

[detail level 1234]

keraflow
activations	Built-in activation functions
backend
common
Backend
tensorflow_backend
TensorflowBackend
theano_backend
TheanoBackend
callbacks
Callback	Abstract base class used to build new callbacks
ModelCheckpoint	Save the model after every epoch
EarlyStopping	Stop training when a monitored quantity has stopped improving
LearningRateScheduler	Learning rate scheduler
constraints	Built-in constrints
MaxNorm	Constrain the weights along an axis pattern to have unit norm
NonNeg	Constrain the weights to be non-negative
UnitNorm	Constrain the weights along an axis pattern to have unit norm
initializations	Built-in initialization functions
layers
base
Kensor	Wrapper for a backend tensor
Layer	Building block of a model
MultiInputLayer	Base class for layer accepting multiple inputs
Input	The entering point of each model
SequentialLayer	Class for making Sequential a Layer
convolution
ConvolutionBase	Base layer for convolution layers
Convolution1D	Convolution layer for convolving (sequence_length, input_dim) inputs
Convolution2D	Convolution layer for convolving (input_depth, input_row, input_col) inputs
Convolution3D	Not implemented yet
PoolingBase	Base layer for pooling layers
Pooling1D	Pooling layer for sub-sampling (sequence_length, input_dim) inputs
Pooling2D	Pooling layer for sub-sampling (input_depth, input_row, input_col) inputs
Pooling3D	Zero-padding layer for (input_depth, input_x, input_y, input_z) inputs
ZeroPaddingBase	Base layer for zero padding layers
ZeroPadding1D	Zero-padding layer for (sequence_length, input_dim) inputs
ZeroPadding2D	Zero-padding layer for (input_depth, input_row, input_col) inputs
ZeroPadding3D	Zero-padding layer for (input_depth, input_x, input_y, input_z) inputs
UnSamplingBase	Base layer for unsampling layers
UnSampling1D	Repeat each temporal step length times along the time axis
UnSampling2D	Unsampling layer for (input_depth, input_row, input_col) inputs
UnSampling3D	Unsampling layer for (input_depth, input_x, input_y, input_z) inputs
core
ExpandDims	Expand dimension of the input tensor
PermuteDims	Permutes the dimensions of the input tensor according to a given pattern
Reshape	Reshapes the input tensor according to a given pattern
Flatten	Flatten the input tensor into 1D
Repeat	Repeat the input tensor n times along given axis
Concatenate	Concatenate multiple input tensors
Lambda	Wrapper for implementating simple inline layer
Activation	Applies an activation function to an output
ElementWiseSum	Reduce multiple input tensors by conducting summation operation
ElementWiseMult	Reduce multiple input tensors by conducting multiplication operation
Dropout	Applies Dropout to the input
Dense	Fully connected layer
Highway	Densely connected highway network
embeddings
Embedding	Vocabulary (row) vectors looking up layer
recurrent
Recurrent	Base class for recurrent layers
SimpleRNN	Fully-connected RNN where the output is to be fed back to input
GRU	Gated Recurrent Unit - Cho et al
LSTM	Long-Short Term Memory unit - Hochreiter 1997
wrappers
TimeDistributed	Wrapper for apply a layer to every temporal slice of an input
models
Model	A model is a directed Kensor graph
Sequential	Model with single input and single output
objectives	Built-in objectives functions
optimizers
Optimizer	Abstract optimizer base class
SGD	Stochastic gradient descent, with support for momentum, learning rate decay, and Nesterov momentum
RMSprop	RMSProp optimizer
Adagrad	Adagrad optimizer
Adadelta	Adadelta optimizer
Adam	Adam optimizer
Adamax	Adamax optimizer from Adam paper's Section 7
regularizers	Built-in regularizers
L1	L1 weight regularization penalty, also known as LASSO
L2	L2 weight regularization penalty, also known as weight decay, or Ridge
L1L2	L1-L2 weight regularization penalty, also known as ElasticNet
utils
exceptions
KeraFlowError
generic_utils	Generic utils for keraflow
user_input_utils
UserInput	Utility class for flexible user input assiging optimizers, regularizers, numpy input..