Typically, we don't care very much about large relative errors in chemical species that have very low abundances. It would be useful to be able to account for this while training/benchmarking by specifying a minimum abundance below which we don't care about the size of the relative error.
Typically, we don't care very much about large relative errors in chemical species that have very low abundances. It would be useful to be able to account for this while training/benchmarking by specifying a minimum abundance below which we don't care about the size of the relative error.