Clean-Clean ER Datasets with FastText embeddings

Please see version 5 for more datasets. This one contains only a print-out of the DBPedia dataset in CSV format.

Each line corresponds to a different entity profile and has the following structure (where n is number of attributes):

numerical_id , uri , n ,  aname_0 , aval_0 , aname_1 , aval_1 ,...aname_n , aval_n  

That is, the separator is "space,space".

Special thanks to moemode for providing the dataset.