Function Reference
Keras Models 


Keras Model 

Keras Model composed of a linear stack of layers 

Replicates a model on different GPUs. 

Print a summary of a Keras model 

Configure a Keras model for training 

Train a Keras model 

Evaluate a Keras model 

Export a Saved Model 

Fits the model on data yielded batchbybatch by a generator. 

Evaluates the model on a data generator. 

Generate predictions from a Keras model 

Generates probability or class probability predictions for the input samples. 

Returns predictions for a single batch of samples. 

Generates predictions for the input samples from a data generator. 

Single gradient update or model evaluation over one batch of samples. 

Retrieves a layer based on either its name (unique) or index. 

Remove the last layer in a model 

Save/Load models using HDF5 files 

Serialize a model to an R object 

Clone a model instance. 

Freeze and unfreeze weights 

Core Layers 

Input layer 

Add a denselyconnected NN layer to an output 

Apply an activation function to an output. 

Applies Dropout to the input. 

Reshapes an output to a certain shape. 

Permute the dimensions of an input according to a given pattern 

Repeats the input n times. 

Wraps arbitrary expression as a layer 

Layer that applies an update to the cost function based input activity. 

Masks a sequence by using a mask value to skip timesteps. 

Flattens an input 

Convolutional Layers 

1D convolution layer (e.g. temporal convolution). 

Transposed 2D convolution layer (sometimes called Deconvolution). 

2D convolution layer (e.g. spatial convolution over images). 

Transposed 3D convolution layer (sometimes called Deconvolution). 

3D convolution layer (e.g. spatial convolution over volumes). 

Convolutional LSTM. 

Depthwise separable 2D convolution. 

Upsampling layer for 1D inputs. 

Upsampling layer for 2D inputs. 

Upsampling layer for 3D inputs. 

Zeropadding layer for 1D input (e.g. temporal sequence). 

Zeropadding layer for 2D input (e.g. picture). 

Zeropadding layer for 3D data (spatial or spatiotemporal). 

Cropping layer for 1D input (e.g. temporal sequence). 

Cropping layer for 2D input (e.g. picture). 

Cropping layer for 3D data (e.g. spatial or spatiotemporal). 

Pooling Layers 

Max pooling operation for temporal data. 

Max pooling operation for spatial data. 

Max pooling operation for 3D data (spatial or spatiotemporal). 

Average pooling for temporal data. 

Average pooling operation for spatial data. 

Average pooling operation for 3D data (spatial or spatiotemporal). 

Global max pooling operation for temporal data. 

Global average pooling operation for temporal data. 

Global max pooling operation for spatial data. 

Global average pooling operation for spatial data. 

Global Max pooling operation for 3D data. 

Global Average pooling operation for 3D data. 

Activation Layers 

Apply an activation function to an output. 

Leaky version of a Rectified Linear Unit. 

Parametric Rectified Linear Unit. 

Thresholded Rectified Linear Unit. 

Exponential Linear Unit. 

Dropout Layers 

Applies Dropout to the input. 

Spatial 1D version of Dropout. 

Spatial 2D version of Dropout. 

Spatial 3D version of Dropout. 

Locallyconnected Layers 

Locallyconnected layer for 1D inputs. 

Locallyconnected layer for 2D inputs. 

Recurrent Layers 

Fullyconnected RNN where the output is to be fed back to input. 

Gated Recurrent Unit  Cho et al. 

Fast GRU implementation backed by CuDNN. 

LongShort Term Memory unit  Hochreiter 1997. 

Fast LSTM implementation backed by CuDNN. 

Embedding Layers 

Turns positive integers (indexes) into dense vectors of fixed size. 

Normalization Layers 

Batch normalization layer (Ioffe and Szegedy, 2014). 

Noise Layers 

Apply additive zerocentered Gaussian noise. 

Apply multiplicative 1centered Gaussian noise. 

Applies Alpha Dropout to the input. 

Merge Layers 

Layer that adds a list of inputs. 

Layer that subtracts two inputs. 

Layer that multiplies (elementwise) a list of inputs. 

Layer that averages a list of inputs. 

Layer that computes the maximum (elementwise) a list of inputs. 

Layer that computes the minimum (elementwise) a list of inputs. 

Layer that concatenates a list of inputs. 

Layer that computes a dot product between samples in two tensors. 

Layer Wrappers 

Apply a layer to every temporal slice of an input. 

Bidirectional wrapper for RNNs. 

Layer Methods 

Layer/Model configuration 

Layer/Model weights as R arrays 


Retrieve tensors for layers with multiple nodes 
Count the total number of scalars composing the weights. 

Reset the states for a layer 

Custom Layers 

Base R6 class for Keras layers 

Create a Keras Layer 

Model Persistence 

Save/Load models using HDF5 files 

Save/Load model weights using HDF5 files 

Serialize a model to an R object 

Layer/Model weights as R arrays 

Layer/Model configuration 

Model configuration as JSON 

Model configuration as YAML 

Datasets 

CIFAR10 small image classification 

CIFAR100 small image classification 

IMDB Movie reviews sentiment classification 

Reuters newswire topics classification 

MNIST database of handwritten digits 

FashionMNIST database of fashion articles 

Boston housing price regression dataset 

Applications 

Xception V1 model for Keras. 

Inception V3 model, with weights pretrained on ImageNet. 


InceptionResNet v2 model, with weights trained on ImageNet 
VGG16 and VGG19 models for Keras. 

ResNet50 model for Keras. 


MobileNet model architecture. 

Instantiates the DenseNet architecture. 

Instantiates a NASNet model. 
Preprocesses a tensor or array encoding a batch of images. 

Decodes the prediction of an ImageNet model. 

Sequence Preprocessing 

Pads each sequence to the same length (length of the longest sequence). 

Generates skipgram word pairs. 

Generates a word rankbased probabilistic sampling table. 

Text Preprocessing 

Text tokenization utility 

Update tokenizer internal vocabulary based on a list of texts or list of sequences. 

Save a text tokenizer to an external file 

Transform each text in texts in a sequence of integers. 

Transforms each text in texts in a sequence of integers. 

Convert a list of texts to a matrix. 

Convert a list of sequences into a matrix. 

Onehot encode a text into a list of word indexes in a vocabulary of size n. 

Converts a text to a sequence of indexes in a fixedsize hashing space. 

Convert text to a sequence of words (or tokens). 

Image Preprocessing 

Loads an image into PIL format. 

3D array representation of images 

Generate minibatches of image data with realtime data augmentation. 

Fit image data generator internal statistics to some sample data. 

Generates batches of augmented/normalized data from image data and labels 

Generates batches of data from images in a directory (with optional augmented/normalized data) 

Retrieve the next item from a generator 

Optimizers 

Stochastic gradient descent optimizer 

RMSProp optimizer 

Adagrad optimizer. 

Adadelta optimizer. 

Adam optimizer 

Adamax optimizer 

Nesterov Adam optimizer 

Callbacks 

Callback that prints metrics to stdout. 

Save the model after every epoch. 

Stop training when a monitored quantity has stopped improving. 

Callback used to stream events to a server. 

Learning rate scheduler. 

TensorBoard basic visualizations 

Reduce learning rate when a metric has stopped improving. 

Callback that terminates training when a NaN loss is encountered. 

Callback that streams epoch results to a csv file 

Create a custom callback 

Base R6 class for Keras callbacks 

Initializers 

Initializer that generates tensors initialized to 0. 

Initializer that generates tensors initialized to 1. 

Initializer that generates tensors initialized to a constant value. 

Initializer that generates tensors with a normal distribution. 

Initializer that generates tensors with a uniform distribution. 

Initializer that generates a truncated normal distribution. 

Initializer capable of adapting its scale to the shape of weights. 

Initializer that generates a random orthogonal matrix. 

Initializer that generates the identity matrix. 

Glorot normal initializer, also called Xavier normal initializer. 

Glorot uniform initializer, also called Xavier uniform initializer. 

He normal initializer. 

He uniform variance scaling initializer. 

LeCun uniform initializer. 

LeCun normal initializer. 

Constraints 


Weight constraints 
Base R6 class for Keras constraints 

Utils 

Plot training history 

Converts a class vector (integers) to binary class matrix. 

Normalize a matrix or ndarray 

Provide a scope with mappings of names to custom objects 

Keras array object 

Representation of HDF5 dataset to be used instead of an R array 

Downloads a file from a URL if it not already in the cache. 

Objects exported from other packages 

Install Keras and the TensorFlow backend 

Check if Keras is Available 

Keras backend tensor engine 

Keras implementation 

Select a Keras implementation and backend 

Losses 


Model loss functions 
Metrics 


Model performance metrics 
Regularizers 

L1 and L2 regularization 

Activations 


Activation functions 
Backend 

Elementwise absolute value. 

Bitwise reduction (logical AND). 

Bitwise reduction (logical OR). 

Creates a 1D tensor containing a sequence of integers. 

Returns the index of the maximum value along an axis. 

Returns the index of the minimum value along an axis. 

Active Keras backend 

Batchwise dot product. 

Turn a nD tensor into a 2D tensor with same 1st dimension. 

Returns the value of more than one tensor variable. 

Applies batch normalization on x given mean, var, beta and gamma. 

Sets the values of many tensor variables at once. 

Adds a bias vector to a tensor. 

Binary crossentropy between an output tensor and a target tensor. 

Cast an array to the default Keras float type. 

Casts a tensor to a different dtype and returns it. 

Categorical crossentropy between an output tensor and a target tensor. 

Destroys the current TF graph and creates a new one. 

Elementwise value clipping. 

Concatenates a list of tensors alongside the specified axis. 

Creates a constant tensor. 

1D convolution. 

2D deconvolution (i.e. transposed convolution). 

2D convolution. 

3D deconvolution (i.e. transposed convolution). 

3D convolution. 

Computes cos of x elementwise. 

Returns the static number of elements in a Keras variable or tensor. 

Runs CTC loss algorithm on each batch element. 

Decodes the output of a softmax. 

Converts CTC labels from dense to sparse. 

Cumulative product of the values in a tensor, alongside the specified axis. 

Cumulative sum of the values in a tensor, alongside the specified axis. 

Depthwise 2D convolution with separable filters. 

Multiplies 2 tensors (and/or variables) and returns a tensor. 

Sets entries in 

Returns the dtype of a Keras tensor or variable, as a string. 

Exponential linear unit. 

Fuzz factor used in numeric expressions. 

Elementwise equality between two tensors. 

Evaluates the value of a variable. 

Elementwise exponential. 

Adds a 1sized dimension at index 

Instantiate an identity matrix and returns it. 

Flatten a tensor. 

Default float type 

Reduce elems using fn to combine them from left to right. 

Reduce elems using fn to combine them from right to left. 

Instantiates a Keras function 

Retrieves the elements of indices 

TF session to be used by the backend. 

Get the uid for the default graph. 

Returns the value of a variable. 

Returns the shape of a variable. 

Returns the gradients of 

Elementwise truth value of (x >= y). 

Elementwise truth value of (x > y). 

Segmentwise linear approximation of sigmoid. 

Returns a tensor with the same content as the input tensor. 

Default image data format convention ('channels_first' or 'channels_last'). 

Selects 

Returns whether the 

Selects 

Returns the shape of tensor or variable as a list of int or NULL entries. 

Returns whether 

Returns whether 

Returns whether a tensor is a sparse tensor. 

Normalizes a tensor wrt the L2 norm alongside the specified axis. 

Returns the learning phase flag. 

Elementwise truth value of (x <= y). 

Elementwise truth value of (x < y). 

Apply 1D conv with unshared weights. 

Apply 2D conv with unshared weights. 

Elementwise log. 

Computes log(sum(exp(elements across dimensions of a tensor))). 

Sets the manual variable initialization flag. 

Map the function fn over the elements elems and return the outputs. 

Maximum value in a tensor. 

Elementwise maximum of two tensors. 

Mean of a tensor, alongside the specified axis. 

Minimum value in a tensor. 

Elementwise minimum of two tensors. 

Compute the moving average of a variable. 

Returns the number of axes in a tensor, as an integer. 

Computes mean and std for batch then apply batch_normalization on batch. 

Elementwise inequality between two tensors. 

Computes the onehot representation of an integer tensor. 

Instantiates an allones variable of the same shape as another tensor. 

Instantiates an allones tensor variable and returns it. 

Permutes axes in a tensor. 

Instantiates a placeholder tensor and returns it. 

2D Pooling. 

3D Pooling. 

Elementwise exponentiation. 

Prints 

Multiplies the values in a tensor, alongside the specified axis. 

Returns a tensor with random binomial distribution of values. 

Instantiates a variable with values drawn from a normal distribution. 

Returns a tensor with normal distribution of values. 

Instantiates a variable with values drawn from a uniform distribution. 

Returns a tensor with uniform distribution of values. 

Rectified linear unit. 

Repeats the elements of a tensor along an axis. 

Repeats a 2D tensor. 

Reset graph identifiers. 

Reshapes a tensor to the specified shape. 

Resizes the images contained in a 4D tensor. 

Resizes the volume contained in a 5D tensor. 

Reverse a tensor along the specified axes. 

Iterates over the time dimension of a tensor 

Elementwise rounding to the closest integer. 

2D convolution with separable filters. 

Sets the learning phase to a fixed value. 

Sets the value of a variable, from an R array. 

Returns the symbolic shape of a tensor or variable. 

Elementwise sigmoid. 

Elementwise sign. 

Computes sin of x elementwise. 

Softmax of a tensor. 

Softplus of a tensor. 

Softsign of a tensor. 

Categorical crossentropy with integer targets. 

Pads the 2nd and 3rd dimensions of a 4D tensor. 

Pads 5D tensor with zeros along the depth, height, width dimensions. 

Elementwise square root. 

Elementwise square. 

Removes a 1dimension from the tensor at index 

Stacks a list of rank 

Standard deviation of a tensor, alongside the specified axis. 

Returns 

Sum of the values in a tensor, alongside the specified axis. 

Switches between two operations depending on a scalar value. 

Elementwise tanh. 

Pads the middle dimension of a 3D tensor. 

Creates a tensor by tiling 

Converts a sparse tensor into a dense tensor and returns it. 

Transposes a tensor and returns it. 

Returns a tensor with truncated random normal distribution of values. 

Update the value of 

Update the value of 

Update the value of 

Variance of a tensor, alongside the specified axis. 

Instantiates a variable and returns it. 

Instantiates an allzeros variable of the same shape as another tensor. 

Instantiates an allzeros variable and returns it. 