> Ocr is well and good, i thought it was mostly solved with tesseract what does this bring? This is specifically for historic documents that tesseract will han

> Ocr is well and good, i thought it was mostly solved with tesseract what does this bring? This is specifically for historic documents that tesseract will handle poorly. It also provides a good interface for retraining models on a specific document set, which will help for documents that are different from the training set.