FUNSD: A Dataset for Form Understanding in Noisy Scanned Documents

Open paper

Benchmark	Model	Rank	Results
semantic-entity-labeling-on-funsd	BERT + MLP Baseline	#18	F1: 57.00