There are 4 levels of OCR quality in wikisource. Can we predict them by looking for common OCR errors?
Description
Description
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | Ladsgroup | T167223 Early June 2017 ORES prod deploy | |||
| Resolved | Halfak | T167196 Implement page_level (OCR) model for frwikisource |