Henning Bonart

Thx @ibaned! I am still working on this issue and will leave this question open for future discussions if it is okay for you.

Well, let me clarify my question: In my simulation code I am doing something like that: ``` old_mesh # generate initial mesh new_mesh = old_mesh old_result # generate initial data...

Mhh, that sounds interesting! What is `set_metric_based_on_result(mesh, new_result);`?

Okay, thx. What about restricting the adaptation to a subarea or submesh?

@ibaned If you have time to spare, could you briefly outline the steps which would be needed to implement a restriction to a submesh?

I will be co-maintainer of this package

I use `LearnerND` with 5 inputs ranging from $\mathcal O(1)$ to $\mathcal O(1\mathrm{e}{-6})$ mapping to one output in the range of $\mathcal O(1\mathrm{e}{-3})$. The function is quite complex and I'm...

> we're not using the value_scale parameter in the loss functions That'd explain my observations, thanks! I Just started sampling a 5D space, before that it was 3D, too. My...

For me too. Edit: permission denied only if inside `import pymc as pm; with pm.Model() as model: [...]`. Edit2: No, something else is going on. Sometimes it failed, and sometimes...