Who can evaluate the dependency tree that StanfordCoreNLP builds?

Question

natural language processing

Who can evaluate the dependency tree that StanfordCoreNLP builds?
[nsubj:pass(oligomictic-2, sandstone-1), dep(quartz-3, oligomictic-2), case(s-4, quartz-3), punct(carton-5, s-4), orphan(type -7, carburizing-5), punct(caruting-5, pin-6), punct(.-8, type-7), root(ROOT-0, .-8)]
Proposal itself:
Oligomictic quartz sandstone with pin cementing type.
It seems that the parser generally gives out some kind of nonsense. Trained on public models of the Russian language.