kadens · PyLaia · Published July 7, 2025
Egerton: English Secretary Hand
Text Recognition
Description
This model transcribes English secretary hand, which was in use during the 16th and 17th centuries. The model was primarily, but not exclusively, trained on equity court material from the National Archives in London, UK. The model has been trained on both clear hands like pleadings and interrogatories and on very difficult hands as found in depositions. It also includes some merchant letters, wills, pages from the London Letter Books, and pages from the Lansdowne collection at the British Library, as well as assorted other material. The model was trained to silently expand most abbreviations. The model was built by a team at Northwestern University Pritzker School of Law in Chicago, Illinois, USA. For more information about Egerton and its transcription conventions see, https://sites.northwestern.edu/Egerton/.
Try this model
Use this modelOpen in Transkribus
Very low error rate2.89% CER
Character Error Rate (CER) measures the percentage of characters incorrectly recognised. Lower is better. This model scored 2.89% on its validation set. As a rule of thumb, a CER below 10% is considered good for most handwritten material. This is a larger model trained on diverse material, which generally makes it more robust across different handwriting styles. That said, larger training sets also make it harder to push the CER down further.
Measured on the model's own validation data. Results on your documents may differ depending on handwriting style, document condition, language, and how closely your material resembles the training data.
Words1,091,795
Lines96,066
Training Pages2,214
Model ID369325