Add the following to the test and check whether the model is created correctly.
e.g.,
s0 = randn(Float32,2)
A = randn(Float32,2,2)
B = randn(Float32,2,2)
Q = randn(Float32,2,2)
R = randn(Float32,2,2)
@test LQDynamicModel(s0,A,B,Q,R,10) isa LQDynamicModel{Float32, Vector{Float32}, SparseArrays.SparseMatrixCSC{Float32, Int64}, SparseArrays.SparseMatrixCSC{Float32, Int64}, Matrix{Float32}}
and do the same for CuArrays. Currently we can't solve it with MadNLP yet, but later we can also test solving it with MadNLP once MadNLP supports single different precisions
Add the following to the test and check whether the model is created correctly.
e.g.,
and do the same for
CuArrays
. Currently we can't solve it with MadNLP yet, but later we can also test solving it with MadNLP once MadNLP supports single different precisions