OCR-to-gold parallel dataset for OCR post-correcting early modern Dutch, originating from EmDComF. Automatically compiled using sentence embeddings and cosine distance.