nvm472001/

cvdataset-layoutlmv3

CV is a collection of receipts. It contains, for each photo about cv personal, a list of OCRs - with the bounding box, text, and class. The goal is to benchmark "key information extraction" - extracting key information from documents https://arxiv.org/abs/2103.14470