Skip to content

v20251125

Choose a tag to compare

View all tags
@n8marti n8marti released this 29 Nov 13:42
· 69 commits to main since this release
20251125
ad5d0ec

This is a traineddata model for Tesseract. Download it to the folder containing other Tesseract .traineddata files, then within Tesseract (CLI) or other frontend GUI app choose "Latin_afr" as the language/script model to use.

Performance

  • Score: 931 (1000 is perfect, Latin model score is 839)
  • Average CER: 3.85%
  • Std. dev. of CER: 3.07%

see Evaluation.md for more

Properties

  • num. of training fonts: 37
  • top layer size: Lfx1536
  • training iterations: 44,900
Assets 3
Loading