zh_core_web_trf / README.md
1
---
2
tags:
3
- spacy
4
- token-classification
5
language:
6
- zh
7
license: mit
8
model-index:
9
- name: zh_core_web_trf
10
  results:
11
  - task:
12
      name: NER
13
      type: token-classification
14
    metrics:
15
    - name: NER Precision
16
      type: precision
17
      value: 0.746365105
18
    - name: NER Recall
19
      type: recall
20
      value: 0.7615384615
21
    - name: NER F Score
22
      type: f_score
23
      value: 0.7538754419
24
  - task:
25
      name: POS
26
      type: token-classification
27
    metrics:
28
    - name: POS Accuracy
29
      type: accuracy
30
      value: 0.9247167985
31
  - task:
32
      name: SENTER
33
      type: token-classification
34
    metrics:
35
    - name: SENTER Precision
36
      type: precision
37
      value: 0.7110739502
38
    - name: SENTER Recall
39
      type: recall
40
      value: 0.6370900616
41
    - name: SENTER F Score
42
      type: f_score
43
      value: 0.67205198
44
  - task:
45
      name: UNLABELED_DEPENDENCIES
46
      type: token-classification
47
    metrics:
48
    - name: Unlabeled Dependencies Accuracy
49
      type: accuracy
50
      value: 0.7683558244
51
  - task:
52
      name: LABELED_DEPENDENCIES
53
      type: token-classification
54
    metrics:
55
    - name: Labeled Dependencies Accuracy
56
      type: accuracy
57
      value: 0.7683558244
58
---
59
### Details: https://spacy.io/models/zh#zh_core_web_trf
60
61
Chinese transformer pipeline (bert-base-chinese). Components: transformer, tagger, parser, ner, attribute_ruler.
62
63
| Feature | Description |
64
| --- | --- |
65
| **Name** | `zh_core_web_trf` |
66
| **Version** | `3.2.0` |
67
| **spaCy** | `>=3.2.0,<3.3.0` |
68
| **Default Pipeline** | `transformer`, `tagger`, `parser`, `attribute_ruler`, `ner` |
69
| **Components** | `transformer`, `tagger`, `parser`, `attribute_ruler`, `ner` |
70
| **Vectors** | 0 keys, 0 unique vectors (0 dimensions) |
71
| **Sources** | [OntoNotes 5](https://catalog.ldc.upenn.edu/LDC2013T19) (Ralph Weischedel, Martha Palmer, Mitchell Marcus, Eduard Hovy, Sameer Pradhan, Lance Ramshaw, Nianwen Xue, Ann Taylor, Jeff Kaufman, Michelle Franchini, Mohammed El-Bachouti, Robert Belvin, Ann Houston)<br />[CoreNLP Universal Dependencies Converter](https://nlp.stanford.edu/software/stanford-dependencies.html) (Stanford NLP Group)<br />[bert-base-chinese](https://huggingface.co/bert-base-chinese) (Hugging Face) |
72
| **License** | `MIT` |
73
| **Author** | [Explosion](https://explosion.ai) |
74
75
### Label Scheme
76
77
<details>
78
79
<summary>View label scheme (99 labels for 3 components)</summary>
80
81
| Component | Labels |
82
| --- | --- |
83
| **`tagger`** | `AD`, `AS`, `BA`, `CC`, `CD`, `CS`, `DEC`, `DEG`, `DER`, `DEV`, `DT`, `ETC`, `FW`, `IJ`, `INF`, `JJ`, `LB`, `LC`, `M`, `MSP`, `NN`, `NR`, `NT`, `OD`, `ON`, `P`, `PN`, `PU`, `SB`, `SP`, `URL`, `VA`, `VC`, `VE`, `VV`, `X` |
84
| **`parser`** | `ROOT`, `acl`, `advcl:loc`, `advmod`, `advmod:dvp`, `advmod:loc`, `advmod:rcomp`, `amod`, `amod:ordmod`, `appos`, `aux:asp`, `aux:ba`, `aux:modal`, `aux:prtmod`, `auxpass`, `case`, `cc`, `ccomp`, `compound:nn`, `compound:vc`, `conj`, `cop`, `dep`, `det`, `discourse`, `dobj`, `etc`, `mark`, `mark:clf`, `name`, `neg`, `nmod`, `nmod:assmod`, `nmod:poss`, `nmod:prep`, `nmod:range`, `nmod:tmod`, `nmod:topic`, `nsubj`, `nsubj:xsubj`, `nsubjpass`, `nummod`, `parataxis:prnmod`, `punct`, `xcomp` |
85
| **`ner`** | `CARDINAL`, `DATE`, `EVENT`, `FAC`, `GPE`, `LANGUAGE`, `LAW`, `LOC`, `MONEY`, `NORP`, `ORDINAL`, `ORG`, `PERCENT`, `PERSON`, `PRODUCT`, `QUANTITY`, `TIME`, `WORK_OF_ART` |
86
87
</details>
88
89
### Accuracy
90
91
| Type | Score |
92
| --- | --- |
93
| `TOKEN_ACC` | 97.88 |
94
| `TOKEN_P` | 94.58 |
95
| `TOKEN_R` | 91.36 |
96
| `TOKEN_F` | 92.94 |
97
| `TAG_ACC` | 92.47 |
98
| `SENTS_P` | 71.11 |
99
| `SENTS_R` | 63.71 |
100
| `SENTS_F` | 67.21 |
101
| `DEP_UAS` | 76.84 |
102
| `DEP_LAS` | 73.07 |
103
| `ENTS_P` | 74.64 |
104
| `ENTS_R` | 76.15 |
105
| `ENTS_F` | 75.39 |