Models

This is a list of trained models for recognition with ocropy.

Latin Scripts

• English Default: http://www.tmbdev.net/en-default.pyrnn.gz (80 MB)
• Fraktur: http://tmbdev.net/ocropy/fraktur.pyrnn.gz (42 MB)
• Fraktur: https://github.com/jze/ocropus-model_fraktur (including training data and a model for CLSTM)
• early 20th century antiqua including south-east European characters: https://github.com/jze/ocropus-model_oesterreich-ungarn (including training data)
• (Old) French: https://github.com/zuphilip/ocropy-french-models (26 MB)
• [Older models by @tmbdev: http://www.tmbdev.net/ocropy/OLD/]

Polytonic Greek Script

• https://github.com/brobertson/ciaconna/tree/master/Classifiers

Cyrillic Scripts

• https://github.com/jze/ocropus-model_cyrillic

Asian Scripts

• Japanese by @isaomatsunami: https://github.com/isaomatsunami/clstm-Japanese (56 MB)

Indic Scripts

• Telugu model & training data: https://github.com/ChillarAnand/likitham

Hume Dialogs text

• Ocropy model for normal text https://github.com/urhub/ocropy/blob/master/models/hume-model-normal-p36-00126000.pyrnn.gz
• Ocropy model for italics text https://github.com/urhub/ocropy/blob/master/models/hume-model-italics--00025000.pyrnn.gz
• Ocropy results, page 230, normal text, https://github.com/urhub/ocropy/blob/master/models/p230_m36n126.pdf
• Ocropy results, page 230, italics text, https://github.com/urhub/ocropy/blob/master/models/p230_mi25.pdf