Automatic Transcription of BnF ms fr 24428 with Transkribus

In this report, we describe the training of a handwritten text recognition (HTR) model in Transkribus for a thirteenth-century medieval French manuscript (Bibliothèque nationale de France, ms. français 24428) using corrected and normalized crowd transcribed training data from the Image du Monde challenge (8-22 January 2021) https://imagedumonde.wordpress.com/. We provide a manuscript description, a statement of transcription principles, comments on the process, brief observations on the performance of the models and a few words on future work. Accompanying the report are four plain text files (two phases of GT used for training and a transcription of the manuscript using two models).