I couldn't find any comparisons with Microsoft's TrOCR model. I guess they are for different purposes. But since you used Florence-2 for OCR, did you compare th

I couldn't find any comparisons with Microsoft's TrOCR model. I guess they are for different purposes. But since you used Florence-2 for OCR, did you compare the two?