API reference#
engine
: Differentiable enginefunctional
: Functions and functionalsactivation
: Activation functionscmass
: Centre of massconnectopy
: Connectopic manifoldscov
: Covariances and correlationsfourier
: Frequency-domain filtergraph
: Graphs and networksinterpolate
: Interpolation and imputationkernel
: Parameterised kernelsmatrix
: Matrix operationsresidualise
semidefinite
: Positive semidefinite conesparse
: Sparse matrix operationssphere
: Spherical coordinatessylo
: Sylo functionsymmap
: Symmetric matrix mapstsconv
: Time series convolutionutils
: Miscellaneouswindow
: Random windowing
init
: Initialisation schemesloss
: Loss and regularisationLoss
: Base class for scalar-valued lossesParameterisedLoss
: Extensible class for custom parameterised lossesMSELoss
: Mean squared errorNormedLoss
: Normed parameter regularisationidentity
: Identity functionzero
: Zero functiondifference
: Elementwise differenceconstraint_violation
: Soft constraintsunilateral_loss
: Unilateral penaltieshinge_loss
: SVM hinge losssmoothness
: Backwards differencesbimodal_symmetric
: Minimal distance from 2 modesdet_gram
: Gramian determinantlog_det_gram
: Gram log-determinant lossentropy
: Categorical entropykl_divergence
: Kullback-Leibler divergencejs_divergence
: Jensen-Shannon divergencebregman_divergence
: Bregman divergencesequilibrium
: Equilibrium losssecond_moment
: Second momentsauto_tol
: Significance tolerancebatch_corr
: Batch-axis correlationqcfc
: QC-FC measures and lossreference_tether
: Spatial tether to reference pointsinterhemispheric_tether
: Inter-hemispheric tethering losscompactness
: Compactnessdispersion
: Vector dispersionmultivariate_kurtosis
: Time series stationarityconnectopy
: Generalised connectopymodularity
: Relaxed modularitysum_scalarise
: Sum scalarisationmean_scalarise
: Mean scalarisationmeansq_scalarise
: Squared mean scalarisationmax_scalarise
: Maximum-value scalarisationnorm_scalarise
: Norm scalarisationvnorm_scalarise
: Vector norm scalarisationwmean_scalarise
: Weighted mean scalarisationselfwmean_scalarise
: Self-weighted mean scalarisationLossApply
: Selectively apply loss to parametersLossScheme
: Scheme for multiple losses
neuro
: Neuroscience data utilitiesnn
: Neural network modules