Just a quick follow-up on this topic, I have been finding that initialising a sparse matrix myself leads to an initialization error:
Rejecting initial value:
Error evaluating the log probability at the initial value.
Exception: cholesky_decompose: Matrix m is not positive definite
I’m guessing this is related to the numerical problem with cholesky_decompose that has been pointed previously https://discourse.mc-stan.org/t/cholesky-factor-corr-initializing-with-invalid-values/3167/6. But I wonder if there is any way around this.