5fa1a76
1
2
The authors introduce a new dataset, PubTables-1M, to benchmark progress in table extraction from unstructured documents, as well as table structure recognition and functional analysis.
