How do these compare to traditional commercial and open source OCR tools? What about things like the Apple Vision APIs?
Yeah I find these discussions without referring to current LTSM model OCRs commonly used like latest version Tesseract strange. I feel like these are the baseline.