Update spaCy pipeline
Browse files- README.md +23 -13
- config.cfg +21 -4
- en_docusco_spacy-any-py3-none-any.whl +2 -2
- meta.json +434 -154
- ner/model +1 -1
- ner/moves +1 -1
- tagger/cfg +280 -0
- tagger/model +3 -0
- tok2vec/model +1 -1
- tokenizer +0 -0
- vocab/strings.json +2 -2
README.md
CHANGED
@@ -13,21 +13,28 @@ model-index:
|
|
13 |
metrics:
|
14 |
- name: NER Precision
|
15 |
type: precision
|
16 |
-
value: 0.
|
17 |
- name: NER Recall
|
18 |
type: recall
|
19 |
-
value: 0.
|
20 |
- name: NER F Score
|
21 |
type: f_score
|
22 |
-
value: 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
23 |
---
|
24 |
| Feature | Description |
|
25 |
| --- | --- |
|
26 |
| **Name** | `en_docusco_spacy` |
|
27 |
-
| **Version** | `0.
|
28 |
-
| **spaCy** | `>=3.
|
29 |
-
| **Default Pipeline** | `tok2vec`, `ner` |
|
30 |
-
| **Components** | `tok2vec`, `ner` |
|
31 |
| **Vectors** | 0 keys, 0 unique vectors (0 dimensions) |
|
32 |
| **Sources** | n/a |
|
33 |
| **License** | n/a |
|
@@ -37,11 +44,12 @@ model-index:
|
|
37 |
|
38 |
<details>
|
39 |
|
40 |
-
<summary>View label scheme (
|
41 |
|
42 |
| Component | Labels |
|
43 |
| --- | --- |
|
44 |
| **`ner`** | `AcademicTerms`, `AcademicWritingMoves`, `Character`, `Citation`, `CitationAuthority`, `CitationHedged`, `ConfidenceHedged`, `ConfidenceHigh`, `ConfidenceLow`, `Contingent`, `Description`, `Facilitate`, `FirstPerson`, `ForceStressed`, `Future`, `InformationChange`, `InformationChangeNegative`, `InformationChangePositive`, `InformationExposition`, `InformationPlace`, `InformationReportVerbs`, `InformationStates`, `InformationTopics`, `Inquiry`, `Interactive`, `MetadiscourseCohesive`, `MetadiscourseInteractive`, `Narrative`, `Negative`, `Positive`, `PublicTerms`, `Reasoning`, `Responsibility`, `Strategic`, `SyntacticComplexity`, `Uncertainty`, `Updates` |
|
|
|
45 |
|
46 |
</details>
|
47 |
|
@@ -49,8 +57,10 @@ model-index:
|
|
49 |
|
50 |
| Type | Score |
|
51 |
| --- | --- |
|
52 |
-
| `ENTS_F` |
|
53 |
-
| `ENTS_P` |
|
54 |
-
| `ENTS_R` | 76
|
55 |
-
| `
|
56 |
-
| `
|
|
|
|
|
|
13 |
metrics:
|
14 |
- name: NER Precision
|
15 |
type: precision
|
16 |
+
value: 0.7498327416
|
17 |
- name: NER Recall
|
18 |
type: recall
|
19 |
+
value: 0.7475641104
|
20 |
- name: NER F Score
|
21 |
type: f_score
|
22 |
+
value: 0.7486967075
|
23 |
+
- task:
|
24 |
+
name: TAG
|
25 |
+
type: token-classification
|
26 |
+
metrics:
|
27 |
+
- name: TAG (XPOS) Accuracy
|
28 |
+
type: accuracy
|
29 |
+
value: 0.9249740665
|
30 |
---
|
31 |
| Feature | Description |
|
32 |
| --- | --- |
|
33 |
| **Name** | `en_docusco_spacy` |
|
34 |
+
| **Version** | `0.2` |
|
35 |
+
| **spaCy** | `>=3.3.0,<3.4.0` |
|
36 |
+
| **Default Pipeline** | `tok2vec`, `ner`, `tagger` |
|
37 |
+
| **Components** | `tok2vec`, `ner`, `tagger` |
|
38 |
| **Vectors** | 0 keys, 0 unique vectors (0 dimensions) |
|
39 |
| **Sources** | n/a |
|
40 |
| **License** | n/a |
|
|
|
44 |
|
45 |
<details>
|
46 |
|
47 |
+
<summary>View label scheme (311 labels for 2 components)</summary>
|
48 |
|
49 |
| Component | Labels |
|
50 |
| --- | --- |
|
51 |
| **`ner`** | `AcademicTerms`, `AcademicWritingMoves`, `Character`, `Citation`, `CitationAuthority`, `CitationHedged`, `ConfidenceHedged`, `ConfidenceHigh`, `ConfidenceLow`, `Contingent`, `Description`, `Facilitate`, `FirstPerson`, `ForceStressed`, `Future`, `InformationChange`, `InformationChangeNegative`, `InformationChangePositive`, `InformationExposition`, `InformationPlace`, `InformationReportVerbs`, `InformationStates`, `InformationTopics`, `Inquiry`, `Interactive`, `MetadiscourseCohesive`, `MetadiscourseInteractive`, `Narrative`, `Negative`, `Positive`, `PublicTerms`, `Reasoning`, `Responsibility`, `Strategic`, `SyntacticComplexity`, `Uncertainty`, `Updates` |
|
52 |
+
| **`tagger`** | `APPGE`, `AT`, `AT1`, `BCL21`, `BCL22`, `CC`, `CCB`, `CS`, `CS21`, `CS22`, `CS31`, `CS32`, `CS33`, `CS41`, `CS42`, `CS43`, `CS44`, `CSA`, `CSN`, `CST`, `CSW`, `CSW31`, `CSW32`, `CSW33`, `DA`, `DA1`, `DA2`, `DAR`, `DAT`, `DB`, `DB2`, `DD`, `DD1`, `DD2`, `DDQ`, `DDQGE`, `DDQGE31`, `DDQGE32`, `DDQGE33`, `DDQV`, `DDQV31`, `DDQV32`, `DDQV33`, `EX`, `FO`, `FU`, `FW`, `GE`, `IF`, `II`, `II21`, `II22`, `II31`, `II32`, `II33`, `II41`, `II42`, `II43`, `II44`, `IO`, `IW`, `JJ`, `JJ21`, `JJ22`, `JJ31`, `JJ32`, `JJ33`, `JJ41`, `JJ42`, `JJ43`, `JJ44`, `JJR`, `JJT`, `JK`, `MC`, `MC1`, `MC2`, `MC221`, `MC222`, `MCMC`, `MD`, `MF`, `ND1`, `NN`, `NN1`, `NN121`, `NN122`, `NN131`, `NN132`, `NN133`, `NN141`, `NN142`, `NN143`, `NN144`, `NN2`, `NN21`, `NN22`, `NN221`, `NN222`, `NN231`, `NN232`, `NN233`, `NN31`, `NN32`, `NN33`, `NNA`, `NNB`, `NNL1`, `NNL2`, `NNO`, `NNO2`, `NNT1`, `NNT131`, `NNT132`, `NNT133`, `NNT2`, `NNU`, `NNU1`, `NNU2`, `NNU21`, `NNU22`, `NP`, `NP1`, `NP2`, `NPD1`, `NPD2`, `NPM1`, `NPM2`, `PN`, `PN1`, `PN121`, `PN122`, `PN21`, `PN22`, `PNQO`, `PNQS`, `PNQS31`, `PNQS32`, `PNQS33`, `PNQV`, `PNQV31`, `PNQV32`, `PNQV33`, `PNX1`, `PPGE`, `PPH1`, `PPHO1`, `PPHO2`, `PPHS1`, `PPHS2`, `PPIO1`, `PPIO2`, `PPIS1`, `PPIS2`, `PPX1`, `PPX121`, `PPX122`, `PPX2`, `PPX221`, `PPX222`, `PPY`, `RA`, `RA21`, `RA22`, `REX`, `REX21`, `REX22`, `REX41`, `REX42`, `REX43`, `REX44`, `RG`, `RG21`, `RG22`, `RG31`, `RG32`, `RG33`, `RG41`, `RG42`, `RG43`, `RG44`, `RGQ`, `RGQV`, `RGQV31`, `RGQV32`, `RGQV33`, `RGR`, `RGT`, `RL`, `RL21`, `RL22`, `RL31`, `RL32`, `RL33`, `RP`, `RPK`, `RR`, `RR21`, `RR22`, `RR31`, `RR32`, `RR33`, `RR41`, `RR42`, `RR43`, `RR44`, `RR51`, `RR52`, `RR53`, `RR54`, `RR55`, `RRQ`, `RRQV`, `RRQV31`, `RRQV32`, `RRQV33`, `RRR`, `RRT`, `RT`, `RT21`, `RT22`, `RT31`, `RT32`, `RT33`, `RT41`, `RT42`, `RT43`, `RT44`, `TO`, `UH`, `UH21`, `UH22`, `UH31`, `UH32`, `UH33`, `VB0`, `VBDR`, `VBDZ`, `VBG`, `VBI`, `VBM`, `VBN`, `VBR`, `VBZ`, `VD0`, `VDD`, `VDG`, `VDI`, `VDN`, `VDZ`, `VH0`, `VHD`, `VHG`, `VHI`, `VHN`, `VHZ`, `VM`, `VM21`, `VM22`, `VMK`, `VV0`, `VVD`, `VVG`, `VVGK`, `VVI`, `VVN`, `VVNK`, `VVZ`, `XX`, `Y`, `ZZ1`, `ZZ2`, `ZZ221`, `ZZ222` |
|
53 |
|
54 |
</details>
|
55 |
|
|
|
57 |
|
58 |
| Type | Score |
|
59 |
| --- | --- |
|
60 |
+
| `ENTS_F` | 74.87 |
|
61 |
+
| `ENTS_P` | 74.98 |
|
62 |
+
| `ENTS_R` | 74.76 |
|
63 |
+
| `TAG_ACC` | 92.50 |
|
64 |
+
| `TOK2VEC_LOSS` | 9919092.94 |
|
65 |
+
| `NER_LOSS` | 6441655.97 |
|
66 |
+
| `TAGGER_LOSS` | 2474292.88 |
|
config.cfg
CHANGED
@@ -1,6 +1,6 @@
|
|
1 |
[paths]
|
2 |
-
train = "
|
3 |
-
dev = "
|
4 |
vectors = null
|
5 |
init_tok2vec = null
|
6 |
|
@@ -10,7 +10,7 @@ seed = 0
|
|
10 |
|
11 |
[nlp]
|
12 |
lang = "en"
|
13 |
-
pipeline = ["tok2vec","ner"]
|
14 |
batch_size = 1000
|
15 |
disabled = []
|
16 |
before_creation = null
|
@@ -41,6 +41,22 @@ nO = null
|
|
41 |
width = ${components.tok2vec.model.encode.width}
|
42 |
upstream = "*"
|
43 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
44 |
[components.tok2vec]
|
45 |
factory = "tok2vec"
|
46 |
|
@@ -123,10 +139,11 @@ eps = 0.00000001
|
|
123 |
learn_rate = 0.001
|
124 |
|
125 |
[training.score_weights]
|
126 |
-
ents_f =
|
127 |
ents_p = 0.0
|
128 |
ents_r = 0.0
|
129 |
ents_per_type = null
|
|
|
130 |
|
131 |
[pretraining]
|
132 |
|
|
|
1 |
[paths]
|
2 |
+
train = "spacy_train.spacy"
|
3 |
+
dev = "spacy_dev.spacy"
|
4 |
vectors = null
|
5 |
init_tok2vec = null
|
6 |
|
|
|
10 |
|
11 |
[nlp]
|
12 |
lang = "en"
|
13 |
+
pipeline = ["tok2vec","ner","tagger"]
|
14 |
batch_size = 1000
|
15 |
disabled = []
|
16 |
before_creation = null
|
|
|
41 |
width = ${components.tok2vec.model.encode.width}
|
42 |
upstream = "*"
|
43 |
|
44 |
+
[components.tagger]
|
45 |
+
factory = "tagger"
|
46 |
+
neg_prefix = "!"
|
47 |
+
overwrite = false
|
48 |
+
scorer = {"@scorers":"spacy.tagger_scorer.v1"}
|
49 |
+
|
50 |
+
[components.tagger.model]
|
51 |
+
@architectures = "spacy.Tagger.v2"
|
52 |
+
nO = null
|
53 |
+
normalize = false
|
54 |
+
|
55 |
+
[components.tagger.model.tok2vec]
|
56 |
+
@architectures = "spacy.Tok2VecListener.v1"
|
57 |
+
width = ${components.tok2vec.model.encode.width}
|
58 |
+
upstream = "*"
|
59 |
+
|
60 |
[components.tok2vec]
|
61 |
factory = "tok2vec"
|
62 |
|
|
|
139 |
learn_rate = 0.001
|
140 |
|
141 |
[training.score_weights]
|
142 |
+
ents_f = 0.5
|
143 |
ents_p = 0.0
|
144 |
ents_r = 0.0
|
145 |
ents_per_type = null
|
146 |
+
tag_acc = 0.5
|
147 |
|
148 |
[pretraining]
|
149 |
|
en_docusco_spacy-any-py3-none-any.whl
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:443d5b1a4188a40e0c1de56e0c9551a97b5b05b013fe995a842ba1840a4dafb0
|
3 |
+
size 6415723
|
meta.json
CHANGED
@@ -1,14 +1,14 @@
|
|
1 |
{
|
2 |
"lang":"en",
|
3 |
"name":"docusco_spacy",
|
4 |
-
"version":"0.
|
5 |
"description":"",
|
6 |
"author":"",
|
7 |
"email":"",
|
8 |
"url":"",
|
9 |
"license":"",
|
10 |
-
"spacy_version":">=3.
|
11 |
-
"spacy_git_version":"
|
12 |
"vectors":{
|
13 |
"width":0,
|
14 |
"vectors":0,
|
@@ -57,212 +57,492 @@
|
|
57 |
"SyntacticComplexity",
|
58 |
"Uncertainty",
|
59 |
"Updates"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
60 |
]
|
61 |
},
|
62 |
"pipeline":[
|
63 |
"tok2vec",
|
64 |
-
"ner"
|
|
|
65 |
],
|
66 |
"components":[
|
67 |
"tok2vec",
|
68 |
-
"ner"
|
|
|
69 |
],
|
70 |
"disabled":[
|
71 |
|
72 |
],
|
73 |
"performance":{
|
74 |
-
"ents_f":0.
|
75 |
-
"ents_p":0.
|
76 |
-
"ents_r":0.
|
77 |
"ents_per_type":{
|
78 |
-
"
|
79 |
-
"p":0.
|
80 |
-
"r":0.
|
81 |
-
"f":0.
|
82 |
-
},
|
83 |
-
"AcademicTerms":{
|
84 |
-
"p":0.6929318188,
|
85 |
-
"r":0.7823984381,
|
86 |
-
"f":0.7349524219
|
87 |
-
},
|
88 |
-
"InformationExposition":{
|
89 |
-
"p":0.8159063958,
|
90 |
-
"r":0.828189449,
|
91 |
-
"f":0.822002039
|
92 |
},
|
93 |
-
"
|
94 |
-
"p":0.
|
95 |
-
"r":0.
|
96 |
-
"f":0.
|
97 |
},
|
98 |
"Character":{
|
99 |
-
"p":0.
|
100 |
-
"r":0.
|
101 |
-
"f":0.
|
102 |
},
|
103 |
-
"
|
104 |
-
"p":0.
|
105 |
-
"r":0.
|
106 |
-
"f":0.
|
|
|
|
|
|
|
|
|
|
|
107 |
},
|
108 |
"Negative":{
|
109 |
-
"p":0.
|
110 |
-
"r":0.
|
111 |
-
"f":0.
|
112 |
},
|
113 |
-
"
|
114 |
-
"p":0.
|
115 |
-
"r":0.
|
116 |
-
"f":0.
|
117 |
},
|
118 |
-
"
|
119 |
-
"p":0.
|
120 |
-
"r":0.
|
121 |
-
"f":0.
|
122 |
},
|
123 |
-
"
|
124 |
-
"p":0.
|
125 |
-
"r":0.
|
126 |
-
"f":0.
|
|
|
|
|
|
|
|
|
|
|
127 |
},
|
128 |
"Interactive":{
|
129 |
-
"p":0.
|
130 |
-
"r":0.
|
131 |
-
"f":0.
|
132 |
},
|
133 |
-
"
|
134 |
-
"p":0.
|
135 |
-
"r":0.
|
136 |
-
"f":0.
|
137 |
},
|
138 |
-
"
|
139 |
-
"p":0.
|
140 |
-
"r":0.
|
141 |
-
"f":0.
|
142 |
},
|
143 |
-
"
|
144 |
-
"p":0.
|
145 |
-
"r":0.
|
146 |
-
"f":0.
|
147 |
},
|
148 |
-
"
|
149 |
-
"p":0.
|
150 |
-
"r":0.
|
151 |
-
"f":0.
|
152 |
},
|
153 |
-
"
|
154 |
-
"p":0.
|
155 |
-
"r":0.
|
156 |
-
"f":0.
|
157 |
},
|
158 |
-
"
|
159 |
-
"p":0.
|
160 |
-
"r":0.
|
161 |
-
"f":0.
|
162 |
},
|
163 |
-
"
|
164 |
-
"p":0.
|
165 |
-
"r":0.
|
166 |
-
"f":0.
|
167 |
},
|
168 |
-
"
|
169 |
-
"p":0.
|
170 |
-
"r":0.
|
171 |
-
"f":0.
|
172 |
},
|
173 |
-
"
|
174 |
-
"p":0.
|
175 |
-
"r":0.
|
176 |
-
"f":0.
|
177 |
},
|
178 |
-
"
|
179 |
-
"p":0.
|
180 |
-
"r":0.
|
181 |
-
"f":0.
|
182 |
},
|
183 |
-
"
|
184 |
-
"p":0.
|
185 |
-
"r":0.
|
186 |
-
"f":0.
|
187 |
},
|
188 |
-
"
|
189 |
-
"p":0.
|
190 |
-
"r":0.
|
191 |
-
"f":0.
|
192 |
},
|
193 |
-
"
|
194 |
-
"p":0.
|
195 |
-
"r":0.
|
196 |
-
"f":0.
|
|
|
|
|
|
|
|
|
|
|
197 |
},
|
198 |
"InformationChange":{
|
199 |
-
"p":0.
|
200 |
-
"r":0.
|
201 |
-
"f":0.
|
202 |
},
|
203 |
-
"
|
204 |
-
"p":0.
|
205 |
-
"r":0.
|
206 |
-
"f":0.
|
207 |
},
|
208 |
-
"
|
209 |
-
"p":0.
|
210 |
-
"r":0.
|
211 |
-
"f":0.
|
212 |
},
|
213 |
-
"
|
214 |
-
"p":0.
|
215 |
-
"r":0.
|
216 |
-
"f":0.
|
217 |
},
|
218 |
-
"
|
219 |
-
"p":0.
|
220 |
-
"r":0.
|
221 |
-
"f":0.
|
222 |
},
|
223 |
-
"
|
224 |
-
"p":0.
|
225 |
-
"r":0.
|
226 |
-
"f":0.
|
227 |
},
|
228 |
-
"
|
229 |
-
"p":0.
|
230 |
-
"r":0.
|
231 |
-
"f":0.
|
232 |
},
|
233 |
-
"
|
234 |
-
"p":0.
|
235 |
-
"r":0.
|
236 |
-
"f":0.
|
237 |
},
|
238 |
-
"
|
239 |
-
"p":0.
|
240 |
-
"r":0.
|
241 |
-
"f":0.
|
242 |
},
|
243 |
"InformationChangePositive":{
|
244 |
-
"p":0.
|
245 |
-
"r":0.
|
246 |
-
"f":0.
|
247 |
-
},
|
248 |
-
"CitationAuthority":{
|
249 |
-
"p":0.7194244604,
|
250 |
-
"r":0.498132005,
|
251 |
-
"f":0.5886681383
|
252 |
},
|
253 |
"ConfidenceLow":{
|
254 |
-
"p":0.
|
255 |
-
"r":0.
|
256 |
-
"f":0.
|
257 |
},
|
258 |
"CitationHedged":{
|
259 |
-
"p":0.
|
260 |
-
"r":0.
|
261 |
-
"f":0.
|
262 |
}
|
263 |
},
|
264 |
-
"
|
265 |
-
"
|
|
|
|
|
266 |
},
|
267 |
"requirements":[
|
268 |
|
|
|
1 |
{
|
2 |
"lang":"en",
|
3 |
"name":"docusco_spacy",
|
4 |
+
"version":"0.2",
|
5 |
"description":"",
|
6 |
"author":"",
|
7 |
"email":"",
|
8 |
"url":"",
|
9 |
"license":"",
|
10 |
+
"spacy_version":">=3.3.0,<3.4.0",
|
11 |
+
"spacy_git_version":"Unknown",
|
12 |
"vectors":{
|
13 |
"width":0,
|
14 |
"vectors":0,
|
|
|
57 |
"SyntacticComplexity",
|
58 |
"Uncertainty",
|
59 |
"Updates"
|
60 |
+
],
|
61 |
+
"tagger":[
|
62 |
+
"APPGE",
|
63 |
+
"AT",
|
64 |
+
"AT1",
|
65 |
+
"BCL21",
|
66 |
+
"BCL22",
|
67 |
+
"CC",
|
68 |
+
"CCB",
|
69 |
+
"CS",
|
70 |
+
"CS21",
|
71 |
+
"CS22",
|
72 |
+
"CS31",
|
73 |
+
"CS32",
|
74 |
+
"CS33",
|
75 |
+
"CS41",
|
76 |
+
"CS42",
|
77 |
+
"CS43",
|
78 |
+
"CS44",
|
79 |
+
"CSA",
|
80 |
+
"CSN",
|
81 |
+
"CST",
|
82 |
+
"CSW",
|
83 |
+
"CSW31",
|
84 |
+
"CSW32",
|
85 |
+
"CSW33",
|
86 |
+
"DA",
|
87 |
+
"DA1",
|
88 |
+
"DA2",
|
89 |
+
"DAR",
|
90 |
+
"DAT",
|
91 |
+
"DB",
|
92 |
+
"DB2",
|
93 |
+
"DD",
|
94 |
+
"DD1",
|
95 |
+
"DD2",
|
96 |
+
"DDQ",
|
97 |
+
"DDQGE",
|
98 |
+
"DDQGE31",
|
99 |
+
"DDQGE32",
|
100 |
+
"DDQGE33",
|
101 |
+
"DDQV",
|
102 |
+
"DDQV31",
|
103 |
+
"DDQV32",
|
104 |
+
"DDQV33",
|
105 |
+
"EX",
|
106 |
+
"FO",
|
107 |
+
"FU",
|
108 |
+
"FW",
|
109 |
+
"GE",
|
110 |
+
"IF",
|
111 |
+
"II",
|
112 |
+
"II21",
|
113 |
+
"II22",
|
114 |
+
"II31",
|
115 |
+
"II32",
|
116 |
+
"II33",
|
117 |
+
"II41",
|
118 |
+
"II42",
|
119 |
+
"II43",
|
120 |
+
"II44",
|
121 |
+
"IO",
|
122 |
+
"IW",
|
123 |
+
"JJ",
|
124 |
+
"JJ21",
|
125 |
+
"JJ22",
|
126 |
+
"JJ31",
|
127 |
+
"JJ32",
|
128 |
+
"JJ33",
|
129 |
+
"JJ41",
|
130 |
+
"JJ42",
|
131 |
+
"JJ43",
|
132 |
+
"JJ44",
|
133 |
+
"JJR",
|
134 |
+
"JJT",
|
135 |
+
"JK",
|
136 |
+
"MC",
|
137 |
+
"MC1",
|
138 |
+
"MC2",
|
139 |
+
"MC221",
|
140 |
+
"MC222",
|
141 |
+
"MCMC",
|
142 |
+
"MD",
|
143 |
+
"MF",
|
144 |
+
"ND1",
|
145 |
+
"NN",
|
146 |
+
"NN1",
|
147 |
+
"NN121",
|
148 |
+
"NN122",
|
149 |
+
"NN131",
|
150 |
+
"NN132",
|
151 |
+
"NN133",
|
152 |
+
"NN141",
|
153 |
+
"NN142",
|
154 |
+
"NN143",
|
155 |
+
"NN144",
|
156 |
+
"NN2",
|
157 |
+
"NN21",
|
158 |
+
"NN22",
|
159 |
+
"NN221",
|
160 |
+
"NN222",
|
161 |
+
"NN231",
|
162 |
+
"NN232",
|
163 |
+
"NN233",
|
164 |
+
"NN31",
|
165 |
+
"NN32",
|
166 |
+
"NN33",
|
167 |
+
"NNA",
|
168 |
+
"NNB",
|
169 |
+
"NNL1",
|
170 |
+
"NNL2",
|
171 |
+
"NNO",
|
172 |
+
"NNO2",
|
173 |
+
"NNT1",
|
174 |
+
"NNT131",
|
175 |
+
"NNT132",
|
176 |
+
"NNT133",
|
177 |
+
"NNT2",
|
178 |
+
"NNU",
|
179 |
+
"NNU1",
|
180 |
+
"NNU2",
|
181 |
+
"NNU21",
|
182 |
+
"NNU22",
|
183 |
+
"NP",
|
184 |
+
"NP1",
|
185 |
+
"NP2",
|
186 |
+
"NPD1",
|
187 |
+
"NPD2",
|
188 |
+
"NPM1",
|
189 |
+
"NPM2",
|
190 |
+
"PN",
|
191 |
+
"PN1",
|
192 |
+
"PN121",
|
193 |
+
"PN122",
|
194 |
+
"PN21",
|
195 |
+
"PN22",
|
196 |
+
"PNQO",
|
197 |
+
"PNQS",
|
198 |
+
"PNQS31",
|
199 |
+
"PNQS32",
|
200 |
+
"PNQS33",
|
201 |
+
"PNQV",
|
202 |
+
"PNQV31",
|
203 |
+
"PNQV32",
|
204 |
+
"PNQV33",
|
205 |
+
"PNX1",
|
206 |
+
"PPGE",
|
207 |
+
"PPH1",
|
208 |
+
"PPHO1",
|
209 |
+
"PPHO2",
|
210 |
+
"PPHS1",
|
211 |
+
"PPHS2",
|
212 |
+
"PPIO1",
|
213 |
+
"PPIO2",
|
214 |
+
"PPIS1",
|
215 |
+
"PPIS2",
|
216 |
+
"PPX1",
|
217 |
+
"PPX121",
|
218 |
+
"PPX122",
|
219 |
+
"PPX2",
|
220 |
+
"PPX221",
|
221 |
+
"PPX222",
|
222 |
+
"PPY",
|
223 |
+
"RA",
|
224 |
+
"RA21",
|
225 |
+
"RA22",
|
226 |
+
"REX",
|
227 |
+
"REX21",
|
228 |
+
"REX22",
|
229 |
+
"REX41",
|
230 |
+
"REX42",
|
231 |
+
"REX43",
|
232 |
+
"REX44",
|
233 |
+
"RG",
|
234 |
+
"RG21",
|
235 |
+
"RG22",
|
236 |
+
"RG31",
|
237 |
+
"RG32",
|
238 |
+
"RG33",
|
239 |
+
"RG41",
|
240 |
+
"RG42",
|
241 |
+
"RG43",
|
242 |
+
"RG44",
|
243 |
+
"RGQ",
|
244 |
+
"RGQV",
|
245 |
+
"RGQV31",
|
246 |
+
"RGQV32",
|
247 |
+
"RGQV33",
|
248 |
+
"RGR",
|
249 |
+
"RGT",
|
250 |
+
"RL",
|
251 |
+
"RL21",
|
252 |
+
"RL22",
|
253 |
+
"RL31",
|
254 |
+
"RL32",
|
255 |
+
"RL33",
|
256 |
+
"RP",
|
257 |
+
"RPK",
|
258 |
+
"RR",
|
259 |
+
"RR21",
|
260 |
+
"RR22",
|
261 |
+
"RR31",
|
262 |
+
"RR32",
|
263 |
+
"RR33",
|
264 |
+
"RR41",
|
265 |
+
"RR42",
|
266 |
+
"RR43",
|
267 |
+
"RR44",
|
268 |
+
"RR51",
|
269 |
+
"RR52",
|
270 |
+
"RR53",
|
271 |
+
"RR54",
|
272 |
+
"RR55",
|
273 |
+
"RRQ",
|
274 |
+
"RRQV",
|
275 |
+
"RRQV31",
|
276 |
+
"RRQV32",
|
277 |
+
"RRQV33",
|
278 |
+
"RRR",
|
279 |
+
"RRT",
|
280 |
+
"RT",
|
281 |
+
"RT21",
|
282 |
+
"RT22",
|
283 |
+
"RT31",
|
284 |
+
"RT32",
|
285 |
+
"RT33",
|
286 |
+
"RT41",
|
287 |
+
"RT42",
|
288 |
+
"RT43",
|
289 |
+
"RT44",
|
290 |
+
"TO",
|
291 |
+
"UH",
|
292 |
+
"UH21",
|
293 |
+
"UH22",
|
294 |
+
"UH31",
|
295 |
+
"UH32",
|
296 |
+
"UH33",
|
297 |
+
"VB0",
|
298 |
+
"VBDR",
|
299 |
+
"VBDZ",
|
300 |
+
"VBG",
|
301 |
+
"VBI",
|
302 |
+
"VBM",
|
303 |
+
"VBN",
|
304 |
+
"VBR",
|
305 |
+
"VBZ",
|
306 |
+
"VD0",
|
307 |
+
"VDD",
|
308 |
+
"VDG",
|
309 |
+
"VDI",
|
310 |
+
"VDN",
|
311 |
+
"VDZ",
|
312 |
+
"VH0",
|
313 |
+
"VHD",
|
314 |
+
"VHG",
|
315 |
+
"VHI",
|
316 |
+
"VHN",
|
317 |
+
"VHZ",
|
318 |
+
"VM",
|
319 |
+
"VM21",
|
320 |
+
"VM22",
|
321 |
+
"VMK",
|
322 |
+
"VV0",
|
323 |
+
"VVD",
|
324 |
+
"VVG",
|
325 |
+
"VVGK",
|
326 |
+
"VVI",
|
327 |
+
"VVN",
|
328 |
+
"VVNK",
|
329 |
+
"VVZ",
|
330 |
+
"XX",
|
331 |
+
"Y",
|
332 |
+
"ZZ1",
|
333 |
+
"ZZ2",
|
334 |
+
"ZZ221",
|
335 |
+
"ZZ222"
|
336 |
]
|
337 |
},
|
338 |
"pipeline":[
|
339 |
"tok2vec",
|
340 |
+
"ner",
|
341 |
+
"tagger"
|
342 |
],
|
343 |
"components":[
|
344 |
"tok2vec",
|
345 |
+
"ner",
|
346 |
+
"tagger"
|
347 |
],
|
348 |
"disabled":[
|
349 |
|
350 |
],
|
351 |
"performance":{
|
352 |
+
"ents_f":0.7486967075,
|
353 |
+
"ents_p":0.7498327416,
|
354 |
+
"ents_r":0.7475641104,
|
355 |
"ents_per_type":{
|
356 |
+
"Narrative":{
|
357 |
+
"p":0.7509423361,
|
358 |
+
"r":0.7103704114,
|
359 |
+
"f":0.7300931537
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
360 |
},
|
361 |
+
"SyntacticComplexity":{
|
362 |
+
"p":0.7886246175,
|
363 |
+
"r":0.8646070539,
|
364 |
+
"f":0.8248697614
|
365 |
},
|
366 |
"Character":{
|
367 |
+
"p":0.785967443,
|
368 |
+
"r":0.7897207251,
|
369 |
+
"f":0.7878396139
|
370 |
},
|
371 |
+
"ConfidenceHigh":{
|
372 |
+
"p":0.7023789438,
|
373 |
+
"r":0.7254298327,
|
374 |
+
"f":0.7137183187
|
375 |
+
},
|
376 |
+
"FirstPerson":{
|
377 |
+
"p":0.8278481013,
|
378 |
+
"r":0.8889566978,
|
379 |
+
"f":0.8573148383
|
380 |
},
|
381 |
"Negative":{
|
382 |
+
"p":0.658557511,
|
383 |
+
"r":0.583953954,
|
384 |
+
"f":0.6190160386
|
385 |
},
|
386 |
+
"Description":{
|
387 |
+
"p":0.6476571898,
|
388 |
+
"r":0.6760073115,
|
389 |
+
"f":0.6615286505
|
390 |
},
|
391 |
+
"InformationExposition":{
|
392 |
+
"p":0.7949913335,
|
393 |
+
"r":0.8201482322,
|
394 |
+
"f":0.8073738649
|
395 |
},
|
396 |
+
"Strategic":{
|
397 |
+
"p":0.6668451911,
|
398 |
+
"r":0.622929668,
|
399 |
+
"f":0.6441397901
|
400 |
+
},
|
401 |
+
"ConfidenceHedged":{
|
402 |
+
"p":0.8015741286,
|
403 |
+
"r":0.8440072161,
|
404 |
+
"f":0.8222435809
|
405 |
},
|
406 |
"Interactive":{
|
407 |
+
"p":0.7874628485,
|
408 |
+
"r":0.8065506265,
|
409 |
+
"f":0.7968924527
|
410 |
},
|
411 |
+
"InformationReportVerbs":{
|
412 |
+
"p":0.7481178396,
|
413 |
+
"r":0.7079138919,
|
414 |
+
"f":0.7274608101
|
415 |
},
|
416 |
+
"ForceStressed":{
|
417 |
+
"p":0.7539183667,
|
418 |
+
"r":0.7242613439,
|
419 |
+
"f":0.7387923478
|
420 |
},
|
421 |
+
"AcademicTerms":{
|
422 |
+
"p":0.7375272228,
|
423 |
+
"r":0.7532883043,
|
424 |
+
"f":0.7453244495
|
425 |
},
|
426 |
+
"MetadiscourseCohesive":{
|
427 |
+
"p":0.904598282,
|
428 |
+
"r":0.9147441302,
|
429 |
+
"f":0.9096429161
|
430 |
},
|
431 |
+
"InformationPlace":{
|
432 |
+
"p":0.8304918033,
|
433 |
+
"r":0.8436303081,
|
434 |
+
"f":0.8370095002
|
435 |
},
|
436 |
+
"Positive":{
|
437 |
+
"p":0.6864356164,
|
438 |
+
"r":0.5833436218,
|
439 |
+
"f":0.6307046557
|
440 |
},
|
441 |
+
"Citation":{
|
442 |
+
"p":0.7408448066,
|
443 |
+
"r":0.7192216767,
|
444 |
+
"f":0.7298731257
|
445 |
},
|
446 |
+
"PublicTerms":{
|
447 |
+
"p":0.7523439934,
|
448 |
+
"r":0.7137098941,
|
449 |
+
"f":0.7325178924
|
450 |
},
|
451 |
+
"CitationAuthority":{
|
452 |
+
"p":0.7350628931,
|
453 |
+
"r":0.4909425046,
|
454 |
+
"f":0.5886982528
|
455 |
},
|
456 |
+
"Reasoning":{
|
457 |
+
"p":0.8164492458,
|
458 |
+
"r":0.7225796753,
|
459 |
+
"f":0.76665178
|
460 |
},
|
461 |
+
"InformationTopics":{
|
462 |
+
"p":0.7416131335,
|
463 |
+
"r":0.7561374001,
|
464 |
+
"f":0.7488048431
|
465 |
},
|
466 |
+
"Responsibility":{
|
467 |
+
"p":0.6843984291,
|
468 |
+
"r":0.4750929368,
|
469 |
+
"f":0.5608543008
|
470 |
},
|
471 |
+
"Updates":{
|
472 |
+
"p":0.7556217827,
|
473 |
+
"r":0.6524994151,
|
474 |
+
"f":0.7002845665
|
475 |
+
},
|
476 |
+
"InformationStates":{
|
477 |
+
"p":0.7582508535,
|
478 |
+
"r":0.8421234057,
|
479 |
+
"f":0.7979893297
|
480 |
},
|
481 |
"InformationChange":{
|
482 |
+
"p":0.6703578762,
|
483 |
+
"r":0.6338672769,
|
484 |
+
"f":0.6516020957
|
485 |
},
|
486 |
+
"Contingent":{
|
487 |
+
"p":0.7394308387,
|
488 |
+
"r":0.7160649222,
|
489 |
+
"f":0.7275603271
|
490 |
},
|
491 |
+
"MetadiscourseInteractive":{
|
492 |
+
"p":0.7715463918,
|
493 |
+
"r":0.6160684886,
|
494 |
+
"f":0.6850970341
|
495 |
},
|
496 |
+
"Future":{
|
497 |
+
"p":0.7021050535,
|
498 |
+
"r":0.6896034438,
|
499 |
+
"f":0.6957980981
|
500 |
},
|
501 |
+
"Inquiry":{
|
502 |
+
"p":0.5940041831,
|
503 |
+
"r":0.5108253947,
|
504 |
+
"f":0.5492836676
|
505 |
},
|
506 |
+
"Uncertainty":{
|
507 |
+
"p":0.6979017644,
|
508 |
+
"r":0.5465919701,
|
509 |
+
"f":0.6130484868
|
510 |
},
|
511 |
+
"AcademicWritingMoves":{
|
512 |
+
"p":0.5667743673,
|
513 |
+
"r":0.42723767,
|
514 |
+
"f":0.4872121282
|
515 |
},
|
516 |
+
"InformationChangeNegative":{
|
517 |
+
"p":0.6899070385,
|
518 |
+
"r":0.4249488753,
|
519 |
+
"f":0.5259427993
|
520 |
},
|
521 |
+
"Facilitate":{
|
522 |
+
"p":0.6855355281,
|
523 |
+
"r":0.5659124447,
|
524 |
+
"f":0.6200067363
|
525 |
},
|
526 |
"InformationChangePositive":{
|
527 |
+
"p":0.7209653092,
|
528 |
+
"r":0.4085470085,
|
529 |
+
"f":0.5215493726
|
|
|
|
|
|
|
|
|
|
|
530 |
},
|
531 |
"ConfidenceLow":{
|
532 |
+
"p":0.751572327,
|
533 |
+
"r":0.4127806563,
|
534 |
+
"f":0.5328874025
|
535 |
},
|
536 |
"CitationHedged":{
|
537 |
+
"p":0.7021276596,
|
538 |
+
"r":0.9328621908,
|
539 |
+
"f":0.8012139605
|
540 |
}
|
541 |
},
|
542 |
+
"tag_acc":0.9249740665,
|
543 |
+
"tok2vec_loss":99190.9293975094,
|
544 |
+
"ner_loss":64416.5596667872,
|
545 |
+
"tagger_loss":24742.9288185574
|
546 |
},
|
547 |
"requirements":[
|
548 |
|
ner/model
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 164952
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d056afe6ac62c1b4455b4418bb51fb2860d465e64b2bb09fd2ce8087b81bf0d6
|
3 |
size 164952
|
ner/moves
CHANGED
@@ -1 +1 @@
|
|
1 |
-
��moves�
|
|
|
1 |
+
��moves�
|
tagger/cfg
ADDED
@@ -0,0 +1,280 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"labels":[
|
3 |
+
"APPGE",
|
4 |
+
"AT",
|
5 |
+
"AT1",
|
6 |
+
"BCL21",
|
7 |
+
"BCL22",
|
8 |
+
"CC",
|
9 |
+
"CCB",
|
10 |
+
"CS",
|
11 |
+
"CS21",
|
12 |
+
"CS22",
|
13 |
+
"CS31",
|
14 |
+
"CS32",
|
15 |
+
"CS33",
|
16 |
+
"CS41",
|
17 |
+
"CS42",
|
18 |
+
"CS43",
|
19 |
+
"CS44",
|
20 |
+
"CSA",
|
21 |
+
"CSN",
|
22 |
+
"CST",
|
23 |
+
"CSW",
|
24 |
+
"CSW31",
|
25 |
+
"CSW32",
|
26 |
+
"CSW33",
|
27 |
+
"DA",
|
28 |
+
"DA1",
|
29 |
+
"DA2",
|
30 |
+
"DAR",
|
31 |
+
"DAT",
|
32 |
+
"DB",
|
33 |
+
"DB2",
|
34 |
+
"DD",
|
35 |
+
"DD1",
|
36 |
+
"DD2",
|
37 |
+
"DDQ",
|
38 |
+
"DDQGE",
|
39 |
+
"DDQGE31",
|
40 |
+
"DDQGE32",
|
41 |
+
"DDQGE33",
|
42 |
+
"DDQV",
|
43 |
+
"DDQV31",
|
44 |
+
"DDQV32",
|
45 |
+
"DDQV33",
|
46 |
+
"EX",
|
47 |
+
"FO",
|
48 |
+
"FU",
|
49 |
+
"FW",
|
50 |
+
"GE",
|
51 |
+
"IF",
|
52 |
+
"II",
|
53 |
+
"II21",
|
54 |
+
"II22",
|
55 |
+
"II31",
|
56 |
+
"II32",
|
57 |
+
"II33",
|
58 |
+
"II41",
|
59 |
+
"II42",
|
60 |
+
"II43",
|
61 |
+
"II44",
|
62 |
+
"IO",
|
63 |
+
"IW",
|
64 |
+
"JJ",
|
65 |
+
"JJ21",
|
66 |
+
"JJ22",
|
67 |
+
"JJ31",
|
68 |
+
"JJ32",
|
69 |
+
"JJ33",
|
70 |
+
"JJ41",
|
71 |
+
"JJ42",
|
72 |
+
"JJ43",
|
73 |
+
"JJ44",
|
74 |
+
"JJR",
|
75 |
+
"JJT",
|
76 |
+
"JK",
|
77 |
+
"MC",
|
78 |
+
"MC1",
|
79 |
+
"MC2",
|
80 |
+
"MC221",
|
81 |
+
"MC222",
|
82 |
+
"MCMC",
|
83 |
+
"MD",
|
84 |
+
"MF",
|
85 |
+
"ND1",
|
86 |
+
"NN",
|
87 |
+
"NN1",
|
88 |
+
"NN121",
|
89 |
+
"NN122",
|
90 |
+
"NN131",
|
91 |
+
"NN132",
|
92 |
+
"NN133",
|
93 |
+
"NN141",
|
94 |
+
"NN142",
|
95 |
+
"NN143",
|
96 |
+
"NN144",
|
97 |
+
"NN2",
|
98 |
+
"NN21",
|
99 |
+
"NN22",
|
100 |
+
"NN221",
|
101 |
+
"NN222",
|
102 |
+
"NN231",
|
103 |
+
"NN232",
|
104 |
+
"NN233",
|
105 |
+
"NN31",
|
106 |
+
"NN32",
|
107 |
+
"NN33",
|
108 |
+
"NNA",
|
109 |
+
"NNB",
|
110 |
+
"NNL1",
|
111 |
+
"NNL2",
|
112 |
+
"NNO",
|
113 |
+
"NNO2",
|
114 |
+
"NNT1",
|
115 |
+
"NNT131",
|
116 |
+
"NNT132",
|
117 |
+
"NNT133",
|
118 |
+
"NNT2",
|
119 |
+
"NNU",
|
120 |
+
"NNU1",
|
121 |
+
"NNU2",
|
122 |
+
"NNU21",
|
123 |
+
"NNU22",
|
124 |
+
"NP",
|
125 |
+
"NP1",
|
126 |
+
"NP2",
|
127 |
+
"NPD1",
|
128 |
+
"NPD2",
|
129 |
+
"NPM1",
|
130 |
+
"NPM2",
|
131 |
+
"PN",
|
132 |
+
"PN1",
|
133 |
+
"PN121",
|
134 |
+
"PN122",
|
135 |
+
"PN21",
|
136 |
+
"PN22",
|
137 |
+
"PNQO",
|
138 |
+
"PNQS",
|
139 |
+
"PNQS31",
|
140 |
+
"PNQS32",
|
141 |
+
"PNQS33",
|
142 |
+
"PNQV",
|
143 |
+
"PNQV31",
|
144 |
+
"PNQV32",
|
145 |
+
"PNQV33",
|
146 |
+
"PNX1",
|
147 |
+
"PPGE",
|
148 |
+
"PPH1",
|
149 |
+
"PPHO1",
|
150 |
+
"PPHO2",
|
151 |
+
"PPHS1",
|
152 |
+
"PPHS2",
|
153 |
+
"PPIO1",
|
154 |
+
"PPIO2",
|
155 |
+
"PPIS1",
|
156 |
+
"PPIS2",
|
157 |
+
"PPX1",
|
158 |
+
"PPX121",
|
159 |
+
"PPX122",
|
160 |
+
"PPX2",
|
161 |
+
"PPX221",
|
162 |
+
"PPX222",
|
163 |
+
"PPY",
|
164 |
+
"RA",
|
165 |
+
"RA21",
|
166 |
+
"RA22",
|
167 |
+
"REX",
|
168 |
+
"REX21",
|
169 |
+
"REX22",
|
170 |
+
"REX41",
|
171 |
+
"REX42",
|
172 |
+
"REX43",
|
173 |
+
"REX44",
|
174 |
+
"RG",
|
175 |
+
"RG21",
|
176 |
+
"RG22",
|
177 |
+
"RG31",
|
178 |
+
"RG32",
|
179 |
+
"RG33",
|
180 |
+
"RG41",
|
181 |
+
"RG42",
|
182 |
+
"RG43",
|
183 |
+
"RG44",
|
184 |
+
"RGQ",
|
185 |
+
"RGQV",
|
186 |
+
"RGQV31",
|
187 |
+
"RGQV32",
|
188 |
+
"RGQV33",
|
189 |
+
"RGR",
|
190 |
+
"RGT",
|
191 |
+
"RL",
|
192 |
+
"RL21",
|
193 |
+
"RL22",
|
194 |
+
"RL31",
|
195 |
+
"RL32",
|
196 |
+
"RL33",
|
197 |
+
"RP",
|
198 |
+
"RPK",
|
199 |
+
"RR",
|
200 |
+
"RR21",
|
201 |
+
"RR22",
|
202 |
+
"RR31",
|
203 |
+
"RR32",
|
204 |
+
"RR33",
|
205 |
+
"RR41",
|
206 |
+
"RR42",
|
207 |
+
"RR43",
|
208 |
+
"RR44",
|
209 |
+
"RR51",
|
210 |
+
"RR52",
|
211 |
+
"RR53",
|
212 |
+
"RR54",
|
213 |
+
"RR55",
|
214 |
+
"RRQ",
|
215 |
+
"RRQV",
|
216 |
+
"RRQV31",
|
217 |
+
"RRQV32",
|
218 |
+
"RRQV33",
|
219 |
+
"RRR",
|
220 |
+
"RRT",
|
221 |
+
"RT",
|
222 |
+
"RT21",
|
223 |
+
"RT22",
|
224 |
+
"RT31",
|
225 |
+
"RT32",
|
226 |
+
"RT33",
|
227 |
+
"RT41",
|
228 |
+
"RT42",
|
229 |
+
"RT43",
|
230 |
+
"RT44",
|
231 |
+
"TO",
|
232 |
+
"UH",
|
233 |
+
"UH21",
|
234 |
+
"UH22",
|
235 |
+
"UH31",
|
236 |
+
"UH32",
|
237 |
+
"UH33",
|
238 |
+
"VB0",
|
239 |
+
"VBDR",
|
240 |
+
"VBDZ",
|
241 |
+
"VBG",
|
242 |
+
"VBI",
|
243 |
+
"VBM",
|
244 |
+
"VBN",
|
245 |
+
"VBR",
|
246 |
+
"VBZ",
|
247 |
+
"VD0",
|
248 |
+
"VDD",
|
249 |
+
"VDG",
|
250 |
+
"VDI",
|
251 |
+
"VDN",
|
252 |
+
"VDZ",
|
253 |
+
"VH0",
|
254 |
+
"VHD",
|
255 |
+
"VHG",
|
256 |
+
"VHI",
|
257 |
+
"VHN",
|
258 |
+
"VHZ",
|
259 |
+
"VM",
|
260 |
+
"VM21",
|
261 |
+
"VM22",
|
262 |
+
"VMK",
|
263 |
+
"VV0",
|
264 |
+
"VVD",
|
265 |
+
"VVG",
|
266 |
+
"VVGK",
|
267 |
+
"VVI",
|
268 |
+
"VVN",
|
269 |
+
"VVNK",
|
270 |
+
"VVZ",
|
271 |
+
"XX",
|
272 |
+
"Y",
|
273 |
+
"ZZ1",
|
274 |
+
"ZZ2",
|
275 |
+
"ZZ221",
|
276 |
+
"ZZ222"
|
277 |
+
],
|
278 |
+
"neg_prefix":"!",
|
279 |
+
"overwrite":false
|
280 |
+
}
|
tagger/model
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:082fdb2ef45d5a5956b0e6b75136ea58e3b385f3b7dd8553f3a4cf2948fa63a9
|
3 |
+
size 106754
|
tok2vec/model
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4443194
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:32b3d9da5e1b1c7d218e12e3ca5acb0ce17645c696ac2385cc56951d9492c2b7
|
3 |
size 4443194
|
tokenizer
CHANGED
The diff for this file is too large to render.
See raw diff
|
|
vocab/strings.json
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c0d0b3cc2098ae111e5a0391aeb7f1b783e85f167d1fdb277933af2fd934308a
|
3 |
+
size 8084734
|