- Dropped support for Ruby < 3.1
- Fixed error with
dup
andclone
- Added
tokenize
andtokenize_file
methods - Added support for untokenized text to text categorization
- Added support for text categorization
- Added support for training binary relation detectors
- Dropped support for Ruby < 2.7
- Added support for training NER models
- Improved ARM detection
- Fixed issue with multibyte characters
- Added ARM shared library for Mac
- Added support for custom tokenization
- Added binary relation detection
- Added
Document
class
- Added shared libraries
- Improved error message when model file does not exist
- First release