Dependency Parsing on deepbiaf NeuroNLP2 implementation (unlabeled metrics)
67.21UCM AccuracyEnsemble
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| EnsembleUpdate Type=deepbiaf to stackptr2022.02 | 67.21 | 2.83 | 8.62 | 92.21 | 1.62 | 20.75 | |
| BCRUpdate Type=deepbiaf to stackptr, Candidate Generation=dropout-p sampling2022.02 | 66.76 | 1.2 | 3.6 | 92.01 | 1.11 | 13.87 | |
| DistillationUpdate Type=deepbiaf to stackptr2022.02 | 66.11 | 3.62 | 10.68 | 91.78 | 2.03 | 24.71 | |
| UntreatedUpdate Type=deepbiaf to stackptr2022.02 | 66.03 | 3.73 | 10.98 | 91.73 | 2.1 | 25.37 | |
| EnsembleUpdate Type=deepbiaf to deepbiaf2022.02 | 64.81 | 2.51 | 7.14 | 92.1 | 1.02 | 12.97 | |
| BCRUpdate Type=deepbiaf to deepbiaf, Candidate Generation=dropout-p sampling2022.02 | 64.36 | 1.12 | 3.14 | 91.78 | 0.84 | 10.21 | |
| DistillationUpdate Type=deepbiaf to deepbiaf2022.02 | 64 | 3.82 | 10.62 | 91.67 | 1.62 | 19.45 | |
| UntreatedUpdate Type=deepbiaf to deepbiaf2022.02 | 63.97 | 3.69 | 10.25 | 91.64 | 1.66 | 19.85 | |
| OldModel Architecture=deepbiaf2022.02 | 63.88 | — | — | 91.76 | — | — | |
| OldModel Architecture=deepbiaf2022.02 | 63.88 | — | — | 91.76 | — | — |