CoreNLP
CoreNLP copied to clipboard
Difference between english_UD to english_SD [Dependency Parser]
What is the difference in terms of quality between the english_UD and english_SD? I see different dependency graph, not only in terms of the labels on the arrows, but also in terms of the arrows themselves.
Is there a difference in quality? Were the models trained on different training set size, or the same training set but with different tagging?