Closed n-kall closed 6 months ago
Ah, yes this happens to me too. Seems to be the print method is failing. The object contains the new fit but must be missing something
I now added a function to create dummy sampler_params
so the brmsfit object no longer gives an error on print.
The issue is that the sampler_params
are stored for each chain, and moment matching merges all chains. Also, as the draws changed, I am not sure that the NUTS parameters are relevant any more anyway. Now in the modified brmsfit, all sampler_params
are just numeric(0)
I now added a function to create dummy
sampler_params
so the brmsfit object no longer gives an error on print.The issue is that the
sampler_params
are stored for each chain, and moment matching merges all chains. Also, as the draws changed, I am not sure that the NUTS parameters are relevant any more anyway. Now in the modified brmsfit, allsampler_params
are justnumeric(0)
Oh that was the missing piece. Thank you for taking care of it! The sampler parameters might be useful for something even after moment matching, but we can set them to 0 for now :+1:
updates to brmsfit method, passes the tests for me. Might be still some work to do to clean it up