Transkribus Community · PyLaia · Published February 19, 2022
Transkribus Print M1
Text Recognition★ Featured
Description
Extended multi-language Transkribus print model, including antiqua and blackletter prints, typewriter, computer print outs and decorative fonts
Includes more languages than print 0.3. The CER in M1 is higher than in 0.3 which is due to a more varied validation set. For languages that were already included in 0.3 the new M1 usually performs equally well or slightly better than 0.3.
Curated by the Transkribus team, this model is occasionally updated with community data for continuous improvement.
Try this model
Use this modelOpen in Transkribus
Very low error rate2.2% CER
Character Error Rate (CER) measures the percentage of characters incorrectly recognised. Lower is better. This model scored 2.2% 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.
Words5,068,310
Lines771,325
Training Pages10,281
Model ID39995