Creator(s)
Project “Bulliot, Bibracte et moi"
Language(s)
Centuries
19th
CER on Validation Set
8.2%
Size (Nr. of Words)
146,466
Model ID
48300
This model has been created within the scope of “Bulliot, Bibracte et moi”, winner of a French Ministry of Culture’s call for Innovative Digital Services. That is a “citizen science” project dedicated to transcript, the notebooks of Jacques-Gabriel Bulliot (1817-1902), who initiated the first excavations at the Bibracte archaeological site between 1867 and 1895 (https://bbm.hypotheses.org). About 147 000 words have been trained for this model and the CER is 8.20%. The CER is relatively high, however, because there are a lot of drawings and plans in which text and dimensions are oriented heterogeneously.
This model was trained in 2021 by Emmanuelle Perrin and Philippe Chassignet (project managers CNRS – Archéorient – UMR 5133, Lyon, France) with the participation of François Chagnot – Jean-Robert Grousset – Sylvie Grousset – Myriam Guillaumet – Isabelle Lagoutte – Marie-Hélène Lidec – Christine Pélardy Seignol – Bruno Ragon – Jean-Claude Seignol – Didier Vernet.
You can use this model to automatically transcribe Handwritten documents with Handwritten Text Recgnition in Transkribus.