(26-10-2026, 09:40 AM)quimqu wrote: You are not allowed to view links. Register or Login to view.Open to questions and doubts.
I see in quire.py this line, to generate a quire trained from the entire VMS without Q13 (M).
# SCRIBE_TRAIN=without_M PYTHONIOENCODING=utf-8 py -3 quire.py M 1 → scribe that has not seen quire M (run calibrate_section_e.py first)
# Outputs in scribe/quire_<Q>/: real_<Q>.txt, scribe_<Q>_all.txt or scribe_<Q>_blind.txt
Is there an option to do the opposite; something like "SCRIBE_TRAIN=with_M" to train on one or several quires? I think it would be more interesting to not mix Currier A and B.
Looking into scribe.py, apparently not.
if _TR.startswith('without_'): return l['quire'] not in _TR[6:].split('+')
Why 6? Should it be 8 (length of without_)?
I suppose I could add:
if _TR.startswith('with_'): return l['quire'] in _TR[5:].split('+')
But calibrate_section_e.py seems to do the training on all sections anyway. I must be missing something.
![[+]](https://www.voynich.ninja/images/collapse_collapsed.png)
