dfish

> 1. Is the optimization result _actually_ done? I could imagine the optimization terminating early or having relatively low default tolerances, which would explain why the result is different at...

My apologies. I think the problem still persists. After restarting my session, it stopped working reliably. I think I may have been working with an old variable. Nonetheless, I think...

The source of the problem is not clear in my opinion. Here is what we observe: when the sample size is small (e.g., 10), the MLEs are highly variable from...

Optimizers accept a function with required inputs. So you would have to modify the example. However, I think you need to use [optim](https://www.rdocumentation.org/packages/stats/versions/3.6.2/topics/optim) with "L-BFGS-B" as the algorithm to make...

Thanks for the suggestion. I came up with my own (see below). My goal was to develop a general function for my package which generates a predictive quantity: ``` function...

@torfjelde, thank you for your explanation. The case you made for using generated quantities with a submodel is compelling. I'll give that a try and if it works out, I...

Awesome! Thanks for suggesting a workaround and fixing the bug.

No problem! I completely understand how that can happen. Thank you for fixing the versions.

Also, it appears that I may have misunderstood the nature of the problem here. If I understand correctly, the deprecation warnings are hidden by default, and fallback on the deprecated...

I think [BayesFlow](https://github.com/stefanradev93/BayesFlow) is another neural network approach that would be good to have.