API documentation

dopt.core.ops.basicContains functions for creating variable nodes and subsequently manipulating their shapes.
dopt.core.ops.mathContains common maths operations.
dopt.core.ops.nnetContains common neural network operations.
dopt.core.ops.randomContains functions for generating random numbers.
dopt.core.gradsContains the automatic differentiation framework.
dopt.core.opsThis package facilitates the construction of various nodes in the operation graph.
dopt.nnet.layers.batchnormContains an implementation of batch normalisation.
dopt.nnet.layers.convContains an implementation of convolutional layers.
dopt.nnet.layers.datasourceAllows one to provide input to a network via a dopt variable.
dopt.nnet.layers.denseContains an implementation of dense (i.e., fully connected) layers.
dopt.nnet.layers.dropoutContains an implementation of dropout.
dopt.nnet.layers.maxpoolContains an implementation of max pooling.
dopt.nnet.layers.reluContains an implementation of the ReLU activation function.
dopt.nnet.layers.softmaxContains an implementation of the softmat activation function.
dopt.nnet.layersContains generic utilities for working with Layer objects.
dopt.nnet.lossesContains some utilities for constructing graphs for common loss functions.
dopt.nnet.networksProvides a useful tools for constructing neural networks.
dopt.nnet.parametersThis module contains methods for initialising the parameters of neural networks.
dopt.online.adamContains an implementation of ADAM that relies on automatic differentiation
dopt.online.amsgradContains an implementation of AMSGrad that relies on automatic differentiation
dopt.online.sgdContains an implementation of stochastic gradient descent that relies on automatic differentiation
dopt.coreThis package contains the framework for constructing and executing operation graphs.
dopt.cpuThis module enables operation graphs to be evaluated using CPU kernels.
dopt.cudaThis is the main interface for the dopt CUDA backend.
dopt.nnetThis package contains a deep learning API backed by dopt.
dopt.onlineThis package contains implementations of common online optimisation algorithms, with a particular bias towards those commonly used in large scale machine learning/deep learning.