flochiff / dahnproject Goto Github PK

It would be good to enhance the metadata of all the documents with an English version so that there are not only encoded in French.

In order to clearly see the three different steps of the transcription (raw transcription, annotation in progress and finalized document), it would be interesting to add a new element in the header.
It would give its status and indicate the level of finalization of the document published on the platform.

The encoding script needs to be improved, following the comments made by @alix-tz

... à partir des données sur d'Estournelles.

• voir quel est le gain au niveau de l'exactitude des segments
• voir quel est le gain en terme de temps d'entrainement

Attention, cela suppose de trouver comment lancer les entraînements sur le cluster en mode gpu...