lengthscale transform #326
Comments
|
I think the ScaleTransform is sufficiently different from the others since it does not operate on vectors necessarily. Regarding ARDTransform, this was discussed in #107. Length scale was discussed in #217 (comment), it seems the issue shouldn't have been closed since this was not resolved/implemented. |
Closed
|
The lengthscale issue is fixed by #336. |
bmharsha
added a commit
to bmharsha/KernelFunctions.jl
that referenced
this issue
Aug 4, 2021
We are encountering following warning during Julia REPL startup if we include `KernelFunctions` in the default SYSIMG, this commit fixes that issue ``` ┌ Warning: Error requiring `PDMats` from `KernelFunctions` │ exception = │ SystemError: opening file "/home/bmharsha/.julia/packages/KernelFunctions/AxuTC/src/matrix/kernelpdmat.jl": No such file or directory │ Stacktrace: │ [1] systemerror(p::String, errno::Int32; extrainfo::Nothing) │ @ Base ./error.jl:168 │ [2] #systemerror#62 │ @ ./error.jl:167 [inlined] │ [3] systemerror │ @ ./error.jl:167 [inlined] │ [4] open(fname::String; lock::Bool, read::Nothing, write::Nothing, create::Nothing, truncate::Nothing, append::Nothing) │ @ Base ./iostream.jl:293 │ [5] open │ @ ./iostream.jl:282 [inlined] │ [6] open(f::Base.var"JuliaGaussianProcesses#326#327"{String}, args::String; kwargs::Base.Iterators.Pairs{Union{}, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}) │ @ Base ./io.jl:328 │ [7] open │ @ ./io.jl:328 [inlined] │ [8] read │ @ ./io.jl:434 [inlined] │ [9] _include(mapexpr::Function, mod::Module, _path::String) │ @ Base ./loading.jl:1166 │ [10] include(mod::Module, _path::String) │ @ Base ./Base.jl:386 │ [11] include(x::String) │ @ KernelFunctions ~/.julia/packages/KernelFunctions/AxuTC/src/KernelFunctions.jl:1 │ [12] top-level scope │ @ ~/.julia/packages/KernelFunctions/AxuTC/src/KernelFunctions.jl:124 │ [13] eval │ @ ./boot.jl:360 [inlined] │ [14] eval │ @ ~/.julia/packages/KernelFunctions/AxuTC/src/KernelFunctions.jl:1 [inlined] │ [15] (::KernelFunctions.var"JuliaGaussianProcesses#209#215")() │ @ KernelFunctions ~/.julia/packages/Requires/7Ncym/src/require.jl:99 │ [16] err(f::Any, listener::Module, modname::String) │ @ Requires ~/.julia/packages/Requires/7Ncym/src/require.jl:47 │ [17] (::KernelFunctions.var"JuliaGaussianProcesses#208#214")() │ @ KernelFunctions ~/.julia/packages/Requires/7Ncym/src/require.jl:98 │ [18] withpath(f::Any, path::String) │ @ Requires ~/.julia/packages/Requires/7Ncym/src/require.jl:37 │ [19] (::KernelFunctions.var"JuliaGaussianProcesses#207#213")() │ @ KernelFunctions ~/.julia/packages/Requires/7Ncym/src/require.jl:97 │ [20] listenpkg(f::Any, pkg::Base.PkgId) │ @ Requires ~/.julia/packages/Requires/7Ncym/src/require.jl:20 │ [21] macro expansion │ @ ~/.julia/packages/Requires/7Ncym/src/require.jl:95 [inlined] │ [22] __init__() │ @ KernelFunctions ~/.julia/packages/KernelFunctions/AxuTC/src/KernelFunctions.jl:123 └ @ Requires ~/.julia/packages/Requires/7Ncym/src/require.jl:49 ```
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
There is currently ScaleTransform, ARDTransform, and LinearTransform. These could all be unified under a LinearTransform with potentially specialised matrices (e.g. a scaled identity matrix or a diagonal matrix for ScaleTransform or ARDTransform, respectively).
Moreover, other kernel/GP packages generally parametrise the kernels with the rather more interpretable lengthscale rather than the inverse lengthscale as KernelFunctions.jl does at the moment: it would be good to have an easy way for users to directly specify a lengthscale (ARD or not).
The text was updated successfully, but these errors were encountered: