2 datasets found

The Penn Treebank and Chinese Treebank are used for constituency and dependency parsing experiments.

The Penn Treebank dataset contains one million words of 1989 Wall Street Journal material annotated in Treebank II style, with 42k sentences of varying lengths.