UD v2.5 benchmarking pipeline for UD_Vietnamese-VTB
| Feature | Description |
|---|---|
| Name | vi_udv25_vietnamesevtb_trf |
| Version | 0.0.1 |
| spaCy | >=3.2.1,<3.3.0 |
| Default Pipeline | experimental_char_ner_tokenizer, transformer, tagger, morphologizer, parser, experimental_edit_tree_lemmatizer |
| Components | experimental_char_ner_tokenizer, transformer, senter, tagger, morphologizer, parser, experimental_edit_tree_lemmatizer |
| Vectors | 0 keys, 0 unique vectors (0 dimensions) |
| Sources | Universal Dependencies v2.5 (Zeman, Daniel; et al.) |
| License | CC BY-SA 4.0 |
| Author | Explosion |
Label Scheme
View label scheme (81 labels for 6 components)
| Component | Labels |
|---|---|
experimental_char_ner_tokenizer |
TOKEN |
senter |
I, S |
tagger |
!, ", ,, -, ., ..., :, ;, ?, @, A, C, CC, E, I, L, LBKT, M, N, NP, Nb, Nc, Np, Nu, Ny, P, R, RBKT, T, V, VP, X, Y, Z |
morphologizer |
POS=NOUN, POS=ADP, POS=X|Polarity=Neg, POS=VERB, POS=ADJ, POS=PUNCT, POS=X, POS=SCONJ, NumType=Card|POS=NUM, POS=DET, POS=CCONJ, POS=PROPN, POS=AUX, POS=PART, POS=INTJ |
parser |
ROOT, advcl, advmod, amod, appos, aux, aux:pass, case, cc, ccomp, compound, conj, cop, csubj, dep, det, discourse, iobj, list, mark, nmod, nsubj, nummod, obj, obl, parataxis, punct, xcomp |
experimental_edit_tree_lemmatizer |
0 |
Accuracy
| Type | Score |
|---|---|
TOKEN_F |
87.90 |
TOKEN_P |
86.84 |
TOKEN_R |
89.00 |
TOKEN_ACC |
98.42 |
SENTS_F |
94.33 |
SENTS_P |
96.23 |
SENTS_R |
92.50 |
TAG_ACC |
88.05 |
POS_ACC |
90.19 |
MORPH_ACC |
96.95 |
DEP_UAS |
68.08 |
DEP_LAS |
60.64 |
LEMMA_ACC |
89.35 |
- Downloads last month
- -
Evaluation results
- TAG (XPOS) Accuracyself-reported0.881
- POS (UPOS) Accuracyself-reported0.902
- Morph (UFeats) Accuracyself-reported0.970
- Lemma Accuracyself-reported0.893
- Unlabeled Attachment Score (UAS)self-reported0.681
- Labeled Attachment Score (LAS)self-reported0.606
- Sentences F-Scoreself-reported0.943