As of now after running simulations, saving the entire simulation object is infeasible, due to size. The solution is to save the result for further processing etc, however when doing this, one becomes unable to see what the config looked like.
The easy quick fix is to save the config as a property within the results object, easiest done when setting up the metric recorder. This might however become redundant, if one does, say a sweep, only changing few parameters at a time.
Maybe a higher layer API could extract the differences in different configurations in a meaningful way to prevent redundant storage of a config.
As of now after running simulations, saving the entire simulation object is infeasible, due to size. The solution is to save the result for further processing etc, however when doing this, one becomes unable to see what the config looked like.
The easy quick fix is to save the config as a property within the results object, easiest done when setting up the metric recorder. This might however become redundant, if one does, say a sweep, only changing few parameters at a time.
Maybe a higher layer API could extract the differences in different configurations in a meaningful way to prevent redundant storage of a config.