Be notified of new releases
Create your free GitHub account today to subscribe to this repository for new releases and build software alongside 28 million developers.Sign up
--fileoption to ocropus-rtrain to read input filenames from a file #275
- Do not add
$datarootdir/ocropusto model search path in windows #268
- Clip exponential in ffunc to avoid overflow #201
- Ignored empty lines in fonts list file #233
- Change checks for write_page_segmentation #220 (allow also small images for segmentation)
- Expand tests for coverage and CI
Testing, continuous integration:
- Test page workflow and confidence measure, update coverage test #145
- circle.yml file for Circle CI testing, run-test-ci #149
- Travis CI builds #37
- Unit tests #209
- Cleanup imports in common.py, lstm.py, extract exceptions #154
- py3k: Use print function instead of statement #155
- py3k: Use new-style exceptions. #175
- Fix behaviour of maxcolseps parameter in ocropus-gpageseg # 172
- Update characters for training #188
- Print summaries to stdout instead of stderr #170
- Connect to CLSTM training and recognition
- Added ocropus-lpred
- Added ocropus-ltrain
- Added ocropus-dewarp
- Added ocropus-linegen
- Added Apache license
- Allow "extract" outside original image bounds #19
- Updated download path for models
- moved ocropus-gtedit back to main directory
A cleaned up version of OCRopy with everything but the new RNN recognizer retired.
Library files that aren't needed anymore have been removed (actually, moved into OLD for the time being).
The installation process has been simplified.
A simple example of training has been added.