Modules

Module containing all publicly available types and procedures

Module containing the activation function setup

Module containing implementation of the leaky ReLU activation function

Module containing implementation of no activation function (identity)

Module containing implementation of the ReLU activation function

Module containing implementation of the sigmoid activation function

Module containing implementation of the swish activation function

Module containing implementation of the activation layer

Module containing implementation of a 1D average pooling layer

Module containing implementation of a 3D average pooling layer

Module containing implementation of 0D and 1D batch normalisation layers

Module containing implementation of 3D batch normalisation layers

Module containing implementation of a concatenation layer

Module containing implementation of a 1D convolutional layer

Module containing implementation of a 3D convolutional layer

Module containing implementation of a 2D dropblock layer

Module containing implementation of a dropout layer

Module containing implementation of a Laplace Neural Operator layer

Module containing implementation of a flattening layer

Module containing implementation of a Graph Neural Operator (GNO) layer

Module containing the implementation of the data initialiser

Module containing the implementation of the Glorot initialiser

Module containing the implementation of the identity initialiser

Module containing the implementation of the Ones initialiser

Module containing procedures for an input layer

Module implementing Kipf & Welling Graph Convolutional Network (GCN)

Module containing loss function implementations

Module containing implementation of a 2D max pooling layer

Module containing functions to compute the accuracy of a model

Module containing custom derived types and interfaces for ATHENA

Module containing the network class used to define a neural network

Module containing procedures for normalising input and output data

Module containing ONNX layer creator functions

Shared utility routines for NOP ONNX export/import.

Module containing implementations of optimisation methods

Module containing implementation of an Orthogonal Neural Operator layer

Module containing implementation of a 2D padding layer

Module containing functions to initialise the random number generator

Module containing regularisation methods

Module containing implementation of a Spectral Filter layer

Athena interface to Weights & Biases (wandb) experiment tracking.