Modules
experimental module
Classes
class AbstractRNNCell: Abstract object representing an RNN cell.
class Activation: Applies an activation function to an output.
class ActivityRegularization: Layer that applies an update to the cost function based input activity.
class Add: Layer that adds a list of inputs.
class AdditiveAttention: Additive attention layer, a.k.a. Bahdanau-style attention.
class AlphaDropout: Applies Alpha Dropout to the input.
class Attention: Dot-product attention layer, a.k.a. Luong-style attention.
class Average: Layer that averages a list of inputs element-wise.
class AveragePooling1D: Average pooling for temporal data.
class AveragePooling2D: Average pooling operation for spatial data.
class AveragePooling3D: Average pooling operation for 3D data (spatial or spatio-temporal).
class AvgPool1D: Average pooling for temporal data.
class AvgPool2D: Average pooling operation for spatial data.
class AvgPool3D: Average pooling operation for 3D data (spatial or spatio-temporal).
class BatchNormalization: Layer that normalizes its inputs.
class Bidirectional: Bidirectional wrapper for RNNs.
class CategoryEncoding: A preprocessing layer which encodes integer features.
class CenterCrop: A preprocessing layer which crops images.
class Concatenate: Layer that concatenates a list of inputs.
class Conv1D: 1D convolution layer (e.g. temporal convolution).
class Conv1DTranspose: Transposed convolution layer (sometimes called Deconvolution).
class Conv2D: 2D convolution layer (e.g. spatial convolution over images).
class Conv2DTranspose: Transposed convolution layer (sometimes called Deconvolution).
class Conv3D: 3D convolution layer (e.g. spatial convolution over volumes).
class Conv3DTranspose: Transposed convolution layer (sometimes called Deconvolution).
class ConvLSTM1D: 1D Convolutional LSTM.
class ConvLSTM2D: 2D Convolutional LSTM.
class ConvLSTM3D: 3D Convolutional LSTM.
class Convolution1D: 1D convolution layer (e.g. temporal convolution).
class Convolution1DTranspose: Transposed convolution layer (sometimes called Deconvolution).
class Convolution2D: 2D convolution layer (e.g. spatial convolution over images).
class Convolution2DTranspose: Transposed convolution layer (sometimes called Deconvolution).
class Convolution3D: 3D convolution layer (e.g. spatial convolution over volumes).
class Convolution3DTranspose: Transposed convolution layer (sometimes called Deconvolution).
class Cropping1D: Cropping layer for 1D input (e.g. temporal sequence).
class Cropping2D: Cropping layer for 2D input (e.g. picture).
class Cropping3D: Cropping layer for 3D data (e.g. spatial or spatio-temporal).
class Dense: Just your regular densely-connected NN layer.
class DenseFeatures: A layer that produces a dense Tensor based on given feature_columns.
class DepthwiseConv1D: Depthwise 1D convolution.
class DepthwiseConv2D: Depthwise 2D convolution.
class Discretization: A preprocessing layer which buckets continuous features by ranges.
class Dot: Layer that computes a dot product between samples in two tensors.
class Dropout: Applies Dropout to the input.
class ELU: Exponential Linear Unit.
class EinsumDense: A layer that uses tf.einsum as the backing computation.
class Embedding: Turns positive integers (indexes) into dense vectors of fixed size.
class Flatten: Flattens the input. Does not affect the batch size.
class GRU: Gated Recurrent Unit - Cho et al. 2014.
class GRUCell: Cell class for the GRU layer.
class GaussianDropout: Apply multiplicative 1-centered Gaussian noise.
class GaussianNoise: Apply additive zero-centered Gaussian noise.
class GlobalAveragePooling1D: Global average pooling operation for temporal data.
class GlobalAveragePooling2D: Global average pooling operation for spatial data.
class GlobalAveragePooling3D: Global Average pooling operation for 3D data.
class GlobalAvgPool1D: Global average pooling operation for temporal data.
class GlobalAvgPool2D: Global average pooling operation for spatial data.
class GlobalAvgPool3D: Global Average pooling operation for 3D data.
class GlobalMaxPool1D: Global max pooling operation for 1D temporal data.
class GlobalMaxPool2D: Global max pooling operation for spatial data.
class GlobalMaxPool3D: Global Max pooling operation for 3D data.
class GlobalMaxPooling1D: Global max pooling operation for 1D temporal data.
class GlobalMaxPooling2D: Global max pooling operation for spatial data.
class GlobalMaxPooling3D: Global Max pooling operation for 3D data.
class GroupNormalization: Group normalization layer.
class HashedCrossing: A preprocessing layer which crosses features using the "hashing trick".
class Hashing: A preprocessing layer which hashes and bins categorical features.
class Identity: Identity layer.
class InputLayer: Layer to be used as an entry point into a Network (a graph of layers).
class InputSpec: Specifies the rank, dtype and shape of every input to a layer.
class IntegerLookup: A preprocessing layer which maps integer features to contiguous ranges.
class LSTM: Long Short-Term Memory layer - Hochreiter 1997.
class LSTMCell: Cell class for the LSTM layer.
class Lambda: Wraps arbitrary expressions as a Layer object.
class Layer: This is the class from which all layers inherit.
class LayerNormalization: Layer normalization layer (Ba et al., 2016).
class LeakyReLU: Leaky version of a Rectified Linear Unit.
class LocallyConnected1D: Locally-connected layer for 1D inputs.
class LocallyConnected2D: Locally-connected layer for 2D inputs.
class Masking: Masks a sequence by using a mask value to skip timesteps.
class MaxPool1D: Max pooling operation for 1D temporal data.
class MaxPool2D: Max pooling operation for 2D spatial data.
class MaxPool3D: Max pooling operation for 3D data (spatial or spatio-temporal).
class MaxPooling1D: Max pooling operation for 1D temporal data.
class MaxPooling2D: Max pooling operation for 2D spatial data.
class MaxPooling3D: Max pooling operation for 3D data (spatial or spatio-temporal).
class Maximum: Layer that computes the maximum (element-wise) a list of inputs.
class Minimum: Layer that computes the minimum (element-wise) a list of inputs.
class MultiHeadAttention: MultiHeadAttention layer.
class Multiply: Layer that multiplies (element-wise) a list of inputs.
class Normalization: A preprocessing layer which normalizes continuous features.
class PReLU: Parametric Rectified Linear Unit.
class Permute: Permutes the dimensions of the input according to a given pattern.
class RNN: Base class for recurrent layers.
class RandomBrightness: A preprocessing layer which randomly adjusts brightness during training.
class RandomContrast: A preprocessing layer which randomly adjusts contrast during training.
class RandomCrop: A preprocessing layer which randomly crops images during training.
class RandomFlip: A preprocessing layer which randomly flips images during training.
class RandomHeight: A preprocessing layer which randomly varies image height during training.
class RandomRotation: A preprocessing layer which randomly rotates images during training.
class RandomTranslation: A preprocessing layer which randomly translates images during training.
class RandomWidth: A preprocessing layer which randomly varies image width during training.
class RandomZoom: A preprocessing layer which randomly zooms images during training.
class ReLU: Rectified Linear Unit activation function.
class RepeatVector: Repeats the input n times.
class Rescaling: A preprocessing layer which rescales input values to a new range.
class Reshape: Layer that reshapes inputs into the given shape.
class Resizing: A preprocessing layer which resizes images.
class SeparableConv1D: Depthwise separable 1D convolution.
class SeparableConv2D: Depthwise separable 2D convolution.
class SeparableConvolution1D: Depthwise separable 1D convolution.
class SeparableConvolution2D: Depthwise separable 2D convolution.
class SimpleRNN: Fully-connected RNN where the output is to be fed back to input.
class SimpleRNNCell: Cell class for SimpleRNN.
class Softmax: Softmax activation function.
class SpatialDropout1D: Spatial 1D version of Dropout.
class SpatialDropout2D: Spatial 2D version of Dropout.
class SpatialDropout3D: Spatial 3D version of Dropout.
class SpectralNormalization: Performs spectral normalization on the weights of a target layer.
class StackedRNNCells: Wrapper allowing a stack of RNN cells to behave as a single cell.
class StringLookup: A preprocessing layer which maps string features to integer indices.
class Subtract: Layer that subtracts two inputs.
class TextVectorization: A preprocessing layer which maps text features to integer sequences.
class ThresholdedReLU: Thresholded Rectified Linear Unit.
class TimeDistributed: This wrapper allows to apply a layer to every temporal slice of an input.
class UnitNormalization: Unit normalization layer.
class UpSampling1D: Upsampling layer for 1D inputs.
class UpSampling2D: Upsampling layer for 2D inputs.
class UpSampling3D: Upsampling layer for 3D inputs.
class Wrapper: Abstract wrapper base class.
class ZeroPadding1D: Zero-padding layer for 1D input (e.g. temporal sequence).
class ZeroPadding2D: Zero-padding layer for 2D input (e.g. picture).
class ZeroPadding3D: Zero-padding layer for 3D data (spatial or spatio-temporal).
Functions
Input(...): Input() is used to instantiate a Keras tensor.
add(...): Functional interface to the tf.keras.layers.Add layer.
average(...): Functional interface to the tf.keras.layers.Average layer.
concatenate(...): Functional interface to the Concatenate layer.
deserialize(...): Instantiates a layer from a config dictionary.
dot(...): Functional interface to the Dot layer.
maximum(...): Functional interface to compute maximum (element-wise) list of inputs.
minimum(...): Functional interface to the Minimum layer.
multiply(...): Functional interface to the Multiply layer.
serialize(...): Serializes a Layer object into a JSON-compatible representation.
subtract(...): Functional interface to the Subtract layer.