The OCR difference is crazy. AI OCR is just isn't there for me and my poor GPU.
Commit
ff41f73
What is your approach to achieving OCR?
What language or tools are you using?
I reverted back to @shipfastlabs Laravel Package, but to make it fast I have a docker container (on my old 2080 super gpu) running paddleocr (i think its called).
Sounds complicated, not so much.
https://github.com/shipfastlabs/parsel
GitHubGitHub - shipfastlabs/parsel: A fast, helpful, and open-source document parser for PHPA fast, helpful, and open-source document parser for PHP - shipfastlabs/parsel
Interesting. I achieved a sort of “manual” OCR in Go using Gosseract (Tesseract in C ported to Go) and noise removal with image-processing libraries
commit ff41f73