Hi @Benedict,
That would be useful indeed, but in the meantime if you are unable to get the specific classifier you need, you can always use the full-length classifier available and play around with the parameters as suggested here. Basically leaving the --p-n-jobs to its default 1 and starting with smaller chunk size using --p-reads-per-batch INTEGER RANGE of say 1000. It’ll take longer but you may not run into the memory issue. Sorry in advance if you’ve already tried these and they haven’t worked.