Skip to content
  • Pricing
susana.filologia · PyLaia · Published September 17, 2026

Spanish_redonda_SO_(HSMS) v1

Text Recognition

Description

This model has been trained to recognize both Roman and Italic typefaces used in Spanish texts printed in the 16th and 17th centuries. The training corpus is based on seven books printed between 1565 and 1630: -Libro de cetreria de caça de açor : en el qual por differente stilo del que tienen los antiguos, que estan hechos, veran (los que a esta caça fueren afficionados) el arte que se ha de tener en el conoscimiento y… (1565) -Primera parte de la Galatea : diuidida en seys libros / cõpuesta por Miguel de Ceruantes… (1585) -Traducion de los libros de Caio Plinio Segundo, de la Historia Natural de los animales / hecha por el Licenciado Geronimo de Huerta ... y anotada por el mesmo ... ; primera parte ... (1599) -El ingenioso hidalgo don Quixote de la Mancha / compuesto por Miguel de Ceruantes Saauedra (1605) -Nouelas exemplares / de Miguel de Ceruantes Saauedra... (1615) -Sitio, naturaleza y propriedades de la ciudad de Mexico aguas y vientos a que esta suieta, y tiempos del año: necessidad de su conocimiento para el exercicio de la medicina, su incertidumbre y difficultad sin el de la astrologia, assi para la curacion como para los prognosticos / por ... Diego Cisneros ... (1618) -Historia de la Vida del Buscon, llamado don Pablos, exemplo de Vagamundos, y espejo de Tacaños / por don Francisco de Queuedo Villegas, Cauallero del Orden de Santiago ... (1630). Ten folios (recto-verso) were selected from five of these editions, whereas twelve folios were chosen from the 1599 and 1618 editions. Finally, the transcription follows the HSMS criteria (http://hispanicseminary.org/manual-en.htm). Therefore, all abbreviations are expanded and enclosed within < > signs, superscript letters are followed by a grave accent, and brackets are replaced with ≺ ≻ signs. However, contrary to the original HSMS guidelines, ç, ñ, and ¶ are transcribed in their original form, rather than converted to their ASCII equivalents (c', n~, and %).

Try this model

Drag an image here

Select a file...

PNG or JPG up to 10 Mb

Wolpi
AI Assistant

By uploading an image, you accept our terms and privacy policy.

Use this modelOpen in Transkribus
Very low error rate2.83% CER

Character Error Rate (CER) measures the percentage of characters incorrectly recognised. Lower is better. This model scored 2.83% on its validation set. As a rule of thumb, a CER below 10% is considered good for most handwritten material.

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.

Words34,666
Lines4,414
Training Pages66
Model ID635449
Languages
LatinCastilian
Centuries
16th c.17th c.