Distributions.MvNormal
— MethodConstruct multivariate normal distribution from GP, facilitate sampling
GPFlux.ArdGaussKernel
— TypeRBF Kernel all parameters are in log-scale
GPFlux.ArdLinearKernel
— TypeLinear Kernel
GPFlux.ArdRQKernel
— TypeRational quardratic kernel
GPFlux.ConstantMean
— TypeConstant mean
GPFlux.GaussProcess
— TypeGauss Process
GPFlux.IsoPeriodKernel
— TypeIso Periodic kernel
GPFlux.NeuralKernelNetwork
— TypeNeural Kernel Network modify https://github.com/FluxML/Flux.jl/tree/master/src/layers/basic.jl Chain
GPFlux.Primitive
— TypePrimitive layer, constituted by basic kernels
GPFlux.ProductCompositeKernel
— Typegeneral composite kernel
GPFlux.Product
— MethodProduct layer
GPFlux.SE_mul_PeriodKernel
— Methodspecial composite kernels
GPFlux.SimpleNeuralNetworkMean
— Methodsimple neural network mean (single layer, linear)
GPFlux.nparams
— Methoddispatch parameters (Array) to a ML model, facilitate use of Optim.jl
GPFlux.positive
— MethodLinear layer: weigths & bias should be positive