Skip to content

OCR models (ocrs), ONNX

Choose a tag to compare

View all tags
@abdolence abdolence released this 01 Nov 11:30
· 42 commits to master since this release
models-ocrs-2
7341620
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Model files for the OCR engine in redacter, in ONNX format, from the ocrs-models project (Robert Knight), used by the ocrs engine. They are copies of the files published at https://ocrs-models.s3-accelerate.amazonaws.com/, which upstream says may change. They replace the .rten files of models-ocrs-1, which redacter 0.21.1 and earlier use.

redacter downloads these files on demand when OCR needs them and checks the size and SHA-256 of each one. There is no need to download them by hand.

File Source Last-Modified Size SHA-256
text-detection.onnx https://ocrs-models.s3-accelerate.amazonaws.com/text-detection.onnx 2026-09-13 08:22:54 GMT 2499479 a917b23dbd9524b465df7e922641b3ff2981623df4ded5a0234004ef2fee7cfe
text-recognition.onnx https://ocrs-models.s3-accelerate.amazonaws.com/text-recognition.onnx 2026-09-13 08:22:52 GMT 9713177 86c145c2edb96c8caed5b1ebb8f44d706408922211451c309c625157dd6061c5

Licence: the models are trained on the HierText dataset and inherit its CC BY-SA 4.0 licence. Attribution: Ocrs models (https://github.com/robertknight/ocrs-models), trained on the HierText dataset by Google Research. Licensed under CC BY-SA 4.0. The ocrs code is MIT OR Apache-2.0. ocrs-models-LICENSE.md and ocrs-models-README.md are copies of the files of ocrs-models at commit a63bee1.

Assets 6
Loading