Transition-based (+BERT + Efficient Training + Effective Encoding)

Reported on 4 benchmarks across 1 task

Note: results are matched by exact model name. Different papers may use the same name for different model variants.