BayKolm/pdf-to-markdown-test-corpus

PDF-to-Markdown Test Corpus 18 small PDFs, each built to break a PDF-to-Markdown converter in one specific way, plus the measured output of one converter against all of them. If you are writing a converter, or choosing one, the hard part is not the happy path. It is knowing what happens when a document has two columns, or a heading that is only bold, or a scanned page in the middle. This corpus is meant to make that testable in about a minute, and to give you somewhere to point… See the full description on the dataset page: https://huggingface.co/datasets/BayKolm/pdf-to-markdown-test-corpus.