Russian Handwriting early 20th century

Free Public AI Model for Handwritten Text Recognition with Transkribus

Russian Handwriting early 20th century

This model is trained on the Russian part of bilingual Evenki/Russian manuscripts by Konstantin Rychkov dated 1911-1913, following pre-reform Cyrillic orthography. These are texts of various genres (folklore, personal narratives, shamanistic rites) collected in different Evenki dialects and translated into Russian. The training set comprises 581 half-pages (59300 words) from several folders of the Rychkov archive. HTR+ engine, 120 epochs.

The model is provided by the INEL project (“Grammatical Descriptions, Corpora, and Language Technology for Indigenous Northern Eurasian Languages”), https://inel.corpora.uni-hamburg.de/.
The archive of Konstantin Rychkov is preserved at the Institute of Oriental Manuscripts RAS, http://www.orientalstudies.ru/. A lower-resolution version of selected manuscripts is accessible at http://www.orientalstudies.ru/rus/index.php?option=com_content&task=view&id=10326&Itemid=149

Model Overview

Name:
Russian-RychkovArchive-v1.0
Creator:
INEL-project
Model ID:
31922
Century:
19th
Languages:
Russian
Script:
Cyrillic alphabet
Engine:
HTR+
Material:
Handwritten
CER on validation set:
4.64 %

Russian-RychkovArchive-v1.0 is freely available to everyone

Get started with Transkribus and use it for your own Material
You can use this model to automatically transcribe Handwritten documents with Handwritten Text Recgnition in Transkribus. This model can be used in the Transkribus Expert Client as well as in Transkribus Lite.
This AI model was trained to automatically convert text from images of historical Cyrillic alphabet documents into editable and searchable text.