Let derivative Kernels take a scalar as `metric` argument as well as list of length `ndim`

karenyyng / shear_gp

code for using GP for inferring the 2D lensing potential. Accompanying gaussian process repo is:

1 stars 0 forks source link

Let derivative Kernels take a scalar as `metric` argument as well as list of length `ndim` #4

Open mdschneider opened 9 years ago

mdschneider commented 9 years ago

The george.Kernel.RadialKernel classes take a metric argument that can be (1) a list of variances for each dimension or (2) a scalar value that is assigned to the variance for all dimensions.

But, the derivative classes (KappaKappa etc.) only allow the first type of argument for metric.