tetrak-easyocr-armenian
Copyright 2026 Stephen Masters, Yvette Mankerian

This product re-exports the recognition network architecture of EasyOCR
(https://github.com/JaidedAI/EasyOCR, Apache License 2.0, JaidedAI).

Trained weights distributed with this project's releases are produced by
tetrak-hy-trainer (https://github.com/scattercode/tetrak-hy-trainer) and
are trained in part on text from Armenian Wikisource's edition of the
Armenian Soviet Encyclopedia (https://hy.wikisource.org/, CC BY-SA 3.0);
per-release attribution and provenance accompany each release.

The word list published beside the weights from v6 on (wordlist.tsv.gz) is
derived from proofread Armenian Wikisource transcripts (CC BY-SA 3.0) and
includes the Nayiri Armenian Lexicon, (c) Serouj Ourishian, licensed under
CC BY 4.0 (http://www.nayiri.com/nayiri-armenian-lexicon), with its word
forms added at the list's minimum count. Because it is derived from those
sources, the word list is licensed under CC BY-SA 4.0
(https://creativecommons.org/licenses/by-sa/4.0/), not under this package's
Apache License 2.0. It is downloaded on request, not distributed in the
package.
