Our current implementations (both 'real' and 'log') give off an awful amount of warnings. Here are a few:
## Rejecting initial value:
Error evaluating the log probability at the initial value.
Exception: lognormal_lpdf: Location parameter[5] is -nan, but must be finite
or
## Exception: ode_rk45: Failed to integrate to next output time (2) in less than max_num_steps steps
which usually lead to
Informational Message: The current Metropolis proposal is about to be rejected because of the
## Chain 1 Exception: ode_rk45: ode parameters and data[1] is inf, but must be finite!
This slows things down by forcing rejections and wasted target and gradient evaluations.
It would be nice to figure out what is going on so we can do away with these warnings and achieve faster computation.
Our current implementations (both 'real' and 'log') give off an awful amount of warnings. Here are a few:
or
which usually lead to
This slows things down by forcing rejections and wasted target and gradient evaluations.
It would be nice to figure out what is going on so we can do away with these warnings and achieve faster computation.