MicrobELP - annotated training sets and unannotated test set documents

This submission contains 3 sets of data: machine_train_with_ref.zip machine_train_without_ref.zip test_without_annotations.zip The training set is given with and without references, and was machine annotated using the pipeline code given in the GitHub. The test set documents are unannotated and can be used for performance benchmarking on Codabench. All files are given in BioC-JSON format obtained using Auto-CORPus.