deeptime.kernels

The kernels package contains kernel implementations which can be used together with kernel methods, such as KernelCCA or KVAD.

Kernel()	The base class of all kernels.
GaussianKernel(sigma[, impl])	Implementation of the Gaussian kernel
GeneralizedGaussianKernel(sigmas)	Implementation of the generalized Gaussian kernel with bandwidth per dimension.
LaplacianKernel(sigma)	Implementation of the Laplacian kernel
PolynomialKernel(degree[, inhomogeneity])	Implementation of the polynomial kernel

PyTorch-ready kernels

TorchGaussianKernel(sigma)

Gaussian kernel which has a dependency to PyTorch and can handle evaluation of Gram matrices as well as (over the batch-size) vectorized evaluations of the kernel.

Utils

is_torch_kernel(kernel)

Tests whether given kernel supports PyTorch.