EC2 Default User
commited on
Commit
•
e3b3969
1
Parent(s):
22d7df0
Update spaCy pipeline
Browse files- README.md +29 -36
- accuracy.json +207 -207
- attribute_ruler/patterns +0 -0
- config.cfg +8 -5
- en_core_web_sm-any-py3-none-any.whl +2 -2
- meta.json +210 -214
- ner/model +2 -2
- parser/model +1 -1
- parser/moves +1 -1
- senter/model +2 -2
- tagger/cfg +1 -0
- tagger/model +2 -2
- tok2vec/model +2 -2
- tokenizer +0 -0
- vocab/strings.json +2 -2
README.md
CHANGED
@@ -14,47 +14,41 @@ model-index:
|
|
14 |
metrics:
|
15 |
- name: NER Precision
|
16 |
type: precision
|
17 |
-
value: 0.
|
18 |
- name: NER Recall
|
19 |
type: recall
|
20 |
-
value: 0.
|
21 |
- name: NER F Score
|
22 |
type: f_score
|
23 |
-
value: 0.
|
24 |
- task:
|
25 |
-
name:
|
26 |
type: token-classification
|
27 |
metrics:
|
28 |
-
- name:
|
29 |
type: accuracy
|
30 |
-
value: 0.
|
31 |
- task:
|
32 |
-
name:
|
33 |
type: token-classification
|
34 |
metrics:
|
35 |
-
- name:
|
36 |
-
type: precision
|
37 |
-
value: 0.9205049471
|
38 |
-
- name: SENTER Recall
|
39 |
-
type: recall
|
40 |
-
value: 0.8899003892
|
41 |
-
- name: SENTER F Score
|
42 |
type: f_score
|
43 |
-
value: 0.
|
44 |
- task:
|
45 |
-
name:
|
46 |
type: token-classification
|
47 |
metrics:
|
48 |
-
- name:
|
49 |
-
type:
|
50 |
-
value: 0.
|
51 |
- task:
|
52 |
-
name:
|
53 |
type: token-classification
|
54 |
metrics:
|
55 |
-
- name:
|
56 |
-
type:
|
57 |
-
value: 0.
|
58 |
---
|
59 |
### Details: https://spacy.io/models/en#en_core_web_sm
|
60 |
|
@@ -63,8 +57,8 @@ English pipeline optimized for CPU. Components: tok2vec, tagger, parser, senter,
|
|
63 |
| Feature | Description |
|
64 |
| --- | --- |
|
65 |
| **Name** | `en_core_web_sm` |
|
66 |
-
| **Version** | `3.
|
67 |
-
| **spaCy** | `>=3.
|
68 |
| **Default Pipeline** | `tok2vec`, `tagger`, `parser`, `attribute_ruler`, `lemmatizer`, `ner` |
|
69 |
| **Components** | `tok2vec`, `tagger`, `parser`, `senter`, `attribute_ruler`, `lemmatizer`, `ner` |
|
70 |
| **Vectors** | 0 keys, 0 unique vectors (0 dimensions) |
|
@@ -76,13 +70,12 @@ English pipeline optimized for CPU. Components: tok2vec, tagger, parser, senter,
|
|
76 |
|
77 |
<details>
|
78 |
|
79 |
-
<summary>View label scheme (
|
80 |
|
81 |
| Component | Labels |
|
82 |
| --- | --- |
|
83 |
| **`tagger`** | `$`, `''`, `,`, `-LRB-`, `-RRB-`, `.`, `:`, `ADD`, `AFX`, `CC`, `CD`, `DT`, `EX`, `FW`, `HYPH`, `IN`, `JJ`, `JJR`, `JJS`, `LS`, `MD`, `NFP`, `NN`, `NNP`, `NNPS`, `NNS`, `PDT`, `POS`, `PRP`, `PRP$`, `RB`, `RBR`, `RBS`, `RP`, `SYM`, `TO`, `UH`, `VB`, `VBD`, `VBG`, `VBN`, `VBP`, `VBZ`, `WDT`, `WP`, `WP$`, `WRB`, `XX`, ```` |
|
84 |
| **`parser`** | `ROOT`, `acl`, `acomp`, `advcl`, `advmod`, `agent`, `amod`, `appos`, `attr`, `aux`, `auxpass`, `case`, `cc`, `ccomp`, `compound`, `conj`, `csubj`, `csubjpass`, `dative`, `dep`, `det`, `dobj`, `expl`, `intj`, `mark`, `meta`, `neg`, `nmod`, `npadvmod`, `nsubj`, `nsubjpass`, `nummod`, `oprd`, `parataxis`, `pcomp`, `pobj`, `poss`, `preconj`, `predet`, `prep`, `prt`, `punct`, `quantmod`, `relcl`, `xcomp` |
|
85 |
-
| **`senter`** | `I`, `S` |
|
86 |
| **`ner`** | `CARDINAL`, `DATE`, `EVENT`, `FAC`, `GPE`, `LANGUAGE`, `LAW`, `LOC`, `MONEY`, `NORP`, `ORDINAL`, `ORG`, `PERCENT`, `PERSON`, `PRODUCT`, `QUANTITY`, `TIME`, `WORK_OF_ART` |
|
87 |
|
88 |
</details>
|
@@ -95,12 +88,12 @@ English pipeline optimized for CPU. Components: tok2vec, tagger, parser, senter,
|
|
95 |
| `TOKEN_P` | 99.57 |
|
96 |
| `TOKEN_R` | 99.58 |
|
97 |
| `TOKEN_F` | 99.57 |
|
98 |
-
| `TAG_ACC` | 97.
|
99 |
-
| `SENTS_P` |
|
100 |
-
| `SENTS_R` |
|
101 |
-
| `SENTS_F` | 90.
|
102 |
-
| `DEP_UAS` | 91.
|
103 |
-
| `DEP_LAS` | 89.
|
104 |
-
| `ENTS_P` |
|
105 |
-
| `ENTS_R` | 83.
|
106 |
-
| `ENTS_F` | 84.
|
|
|
14 |
metrics:
|
15 |
- name: NER Precision
|
16 |
type: precision
|
17 |
+
value: 0.8508041869
|
18 |
- name: NER Recall
|
19 |
type: recall
|
20 |
+
value: 0.8344851763
|
21 |
- name: NER F Score
|
22 |
type: f_score
|
23 |
+
value: 0.8425656714
|
24 |
- task:
|
25 |
+
name: TAG
|
26 |
type: token-classification
|
27 |
metrics:
|
28 |
+
- name: TAG (XPOS) Accuracy
|
29 |
type: accuracy
|
30 |
+
value: 0.9726545475
|
31 |
- task:
|
32 |
+
name: UNLABELED_DEPENDENCIES
|
33 |
type: token-classification
|
34 |
metrics:
|
35 |
+
- name: Unlabeled Attachment Score (UAS)
|
|
|
|
|
|
|
|
|
|
|
|
|
36 |
type: f_score
|
37 |
+
value: 0.9180803841
|
38 |
- task:
|
39 |
+
name: LABELED_DEPENDENCIES
|
40 |
type: token-classification
|
41 |
metrics:
|
42 |
+
- name: Labeled Attachment Score (LAS)
|
43 |
+
type: f_score
|
44 |
+
value: 0.8996666011
|
45 |
- task:
|
46 |
+
name: SENTS
|
47 |
type: token-classification
|
48 |
metrics:
|
49 |
+
- name: Sentences F-Score
|
50 |
+
type: f_score
|
51 |
+
value: 0.9060200669
|
52 |
---
|
53 |
### Details: https://spacy.io/models/en#en_core_web_sm
|
54 |
|
|
|
57 |
| Feature | Description |
|
58 |
| --- | --- |
|
59 |
| **Name** | `en_core_web_sm` |
|
60 |
+
| **Version** | `3.3.0` |
|
61 |
+
| **spaCy** | `>=3.3.0.dev0,<3.4.0` |
|
62 |
| **Default Pipeline** | `tok2vec`, `tagger`, `parser`, `attribute_ruler`, `lemmatizer`, `ner` |
|
63 |
| **Components** | `tok2vec`, `tagger`, `parser`, `senter`, `attribute_ruler`, `lemmatizer`, `ner` |
|
64 |
| **Vectors** | 0 keys, 0 unique vectors (0 dimensions) |
|
|
|
70 |
|
71 |
<details>
|
72 |
|
73 |
+
<summary>View label scheme (112 labels for 3 components)</summary>
|
74 |
|
75 |
| Component | Labels |
|
76 |
| --- | --- |
|
77 |
| **`tagger`** | `$`, `''`, `,`, `-LRB-`, `-RRB-`, `.`, `:`, `ADD`, `AFX`, `CC`, `CD`, `DT`, `EX`, `FW`, `HYPH`, `IN`, `JJ`, `JJR`, `JJS`, `LS`, `MD`, `NFP`, `NN`, `NNP`, `NNPS`, `NNS`, `PDT`, `POS`, `PRP`, `PRP$`, `RB`, `RBR`, `RBS`, `RP`, `SYM`, `TO`, `UH`, `VB`, `VBD`, `VBG`, `VBN`, `VBP`, `VBZ`, `WDT`, `WP`, `WP$`, `WRB`, `XX`, ```` |
|
78 |
| **`parser`** | `ROOT`, `acl`, `acomp`, `advcl`, `advmod`, `agent`, `amod`, `appos`, `attr`, `aux`, `auxpass`, `case`, `cc`, `ccomp`, `compound`, `conj`, `csubj`, `csubjpass`, `dative`, `dep`, `det`, `dobj`, `expl`, `intj`, `mark`, `meta`, `neg`, `nmod`, `npadvmod`, `nsubj`, `nsubjpass`, `nummod`, `oprd`, `parataxis`, `pcomp`, `pobj`, `poss`, `preconj`, `predet`, `prep`, `prt`, `punct`, `quantmod`, `relcl`, `xcomp` |
|
|
|
79 |
| **`ner`** | `CARDINAL`, `DATE`, `EVENT`, `FAC`, `GPE`, `LANGUAGE`, `LAW`, `LOC`, `MONEY`, `NORP`, `ORDINAL`, `ORG`, `PERCENT`, `PERSON`, `PRODUCT`, `QUANTITY`, `TIME`, `WORK_OF_ART` |
|
80 |
|
81 |
</details>
|
|
|
88 |
| `TOKEN_P` | 99.57 |
|
89 |
| `TOKEN_R` | 99.58 |
|
90 |
| `TOKEN_F` | 99.57 |
|
91 |
+
| `TAG_ACC` | 97.27 |
|
92 |
+
| `SENTS_P` | 91.89 |
|
93 |
+
| `SENTS_R` | 89.35 |
|
94 |
+
| `SENTS_F` | 90.60 |
|
95 |
+
| `DEP_UAS` | 91.81 |
|
96 |
+
| `DEP_LAS` | 89.97 |
|
97 |
+
| `ENTS_P` | 85.08 |
|
98 |
+
| `ENTS_R` | 83.45 |
|
99 |
+
| `ENTS_F` | 84.26 |
|
accuracy.json
CHANGED
@@ -1,330 +1,330 @@
|
|
1 |
{
|
2 |
-
"token_acc": 0.
|
3 |
-
"token_p": 0.
|
4 |
-
"token_r": 0.
|
5 |
-
"token_f": 0.
|
6 |
-
"tag_acc": 0.
|
7 |
-
"sents_p": 0.
|
8 |
-
"sents_r": 0.
|
9 |
-
"sents_f": 0.
|
10 |
-
"dep_uas": 0.
|
11 |
-
"dep_las": 0.
|
12 |
"dep_las_per_type": {
|
13 |
"prep": {
|
14 |
-
"p": 0.
|
15 |
-
"r": 0.
|
16 |
-
"f": 0.
|
17 |
},
|
18 |
"det": {
|
19 |
-
"p": 0.
|
20 |
"r": 0.9787164642,
|
21 |
-
"f": 0.
|
22 |
},
|
23 |
"pobj": {
|
24 |
-
"p": 0.
|
25 |
-
"r": 0.
|
26 |
-
"f": 0.
|
27 |
},
|
28 |
"nsubj": {
|
29 |
-
"p": 0.
|
30 |
-
"r": 0.
|
31 |
-
"f": 0.
|
32 |
},
|
33 |
"aux": {
|
34 |
-
"p": 0.
|
35 |
-
"r": 0.
|
36 |
-
"f": 0.
|
37 |
},
|
38 |
"advmod": {
|
39 |
-
"p": 0.
|
40 |
-
"r": 0.
|
41 |
-
"f": 0.
|
42 |
},
|
43 |
"relcl": {
|
44 |
-
"p": 0.
|
45 |
-
"r": 0.
|
46 |
-
"f": 0.
|
47 |
},
|
48 |
"root": {
|
49 |
-
"p": 0.
|
50 |
-
"r": 0.
|
51 |
-
"f": 0.
|
52 |
},
|
53 |
"xcomp": {
|
54 |
-
"p": 0.
|
55 |
-
"r": 0.
|
56 |
-
"f": 0.
|
57 |
},
|
58 |
"amod": {
|
59 |
-
"p": 0.
|
60 |
-
"r": 0.
|
61 |
-
"f": 0.
|
62 |
},
|
63 |
"compound": {
|
64 |
-
"p": 0.
|
65 |
-
"r": 0.
|
66 |
-
"f": 0.
|
67 |
},
|
68 |
"poss": {
|
69 |
-
"p": 0.
|
70 |
-
"r": 0.
|
71 |
-
"f": 0.
|
72 |
},
|
73 |
"ccomp": {
|
74 |
-
"p": 0.
|
75 |
-
"r": 0.
|
76 |
-
"f": 0.
|
77 |
},
|
78 |
"attr": {
|
79 |
-
"p": 0.
|
80 |
-
"r": 0.
|
81 |
-
"f": 0.
|
82 |
},
|
83 |
"case": {
|
84 |
-
"p": 0.
|
85 |
-
"r": 0.
|
86 |
-
"f": 0.
|
87 |
},
|
88 |
"mark": {
|
89 |
-
"p": 0.
|
90 |
-
"r": 0.
|
91 |
-
"f": 0.
|
92 |
},
|
93 |
"intj": {
|
94 |
-
"p": 0.
|
95 |
-
"r": 0.
|
96 |
-
"f": 0.
|
97 |
},
|
98 |
"advcl": {
|
99 |
-
"p": 0.
|
100 |
-
"r": 0.
|
101 |
-
"f": 0.
|
102 |
},
|
103 |
"cc": {
|
104 |
-
"p": 0.
|
105 |
-
"r": 0.
|
106 |
-
"f": 0.
|
107 |
},
|
108 |
"neg": {
|
109 |
-
"p": 0.
|
110 |
-
"r": 0.
|
111 |
-
"f": 0.
|
112 |
},
|
113 |
"conj": {
|
114 |
-
"p": 0.
|
115 |
-
"r": 0.
|
116 |
-
"f": 0.
|
117 |
},
|
118 |
"nsubjpass": {
|
119 |
-
"p": 0.
|
120 |
-
"r": 0.
|
121 |
-
"f": 0.
|
122 |
},
|
123 |
"auxpass": {
|
124 |
-
"p": 0.
|
125 |
-
"r": 0.
|
126 |
-
"f": 0.
|
127 |
},
|
128 |
"dobj": {
|
129 |
-
"p": 0.
|
130 |
-
"r": 0.
|
131 |
-
"f": 0.
|
132 |
},
|
133 |
"nummod": {
|
134 |
-
"p": 0.
|
135 |
-
"r": 0.
|
136 |
-
"f": 0.
|
137 |
},
|
138 |
"npadvmod": {
|
139 |
-
"p": 0.
|
140 |
-
"r": 0.
|
141 |
-
"f": 0.
|
142 |
},
|
143 |
"prt": {
|
144 |
-
"p": 0.
|
145 |
-
"r": 0.
|
146 |
-
"f": 0.
|
147 |
},
|
148 |
"pcomp": {
|
149 |
-
"p": 0.
|
150 |
-
"r": 0.
|
151 |
-
"f": 0.
|
152 |
},
|
153 |
"expl": {
|
154 |
-
"p": 0.
|
155 |
-
"r": 0.
|
156 |
-
"f": 0.
|
157 |
},
|
158 |
"acl": {
|
159 |
-
"p": 0.
|
160 |
-
"r": 0.
|
161 |
-
"f": 0.
|
162 |
},
|
163 |
"agent": {
|
164 |
-
"p": 0.
|
165 |
-
"r": 0.
|
166 |
-
"f": 0.
|
167 |
},
|
168 |
"dative": {
|
169 |
-
"p": 0.
|
170 |
-
"r": 0.
|
171 |
-
"f": 0.
|
172 |
},
|
173 |
"acomp": {
|
174 |
-
"p": 0.
|
175 |
-
"r": 0.
|
176 |
-
"f": 0.
|
177 |
},
|
178 |
"dep": {
|
179 |
-
"p": 0.
|
180 |
-
"r": 0.
|
181 |
-
"f": 0.
|
182 |
},
|
183 |
"csubj": {
|
184 |
-
"p": 0.
|
185 |
-
"r": 0.
|
186 |
-
"f": 0.
|
187 |
},
|
188 |
"quantmod": {
|
189 |
-
"p": 0.
|
190 |
-
"r": 0.
|
191 |
-
"f": 0.
|
192 |
},
|
193 |
"nmod": {
|
194 |
-
"p": 0.
|
195 |
-
"r": 0.
|
196 |
-
"f": 0.
|
197 |
},
|
198 |
"appos": {
|
199 |
-
"p": 0.
|
200 |
-
"r": 0.
|
201 |
-
"f": 0.
|
202 |
},
|
203 |
"predet": {
|
204 |
-
"p": 0.
|
205 |
-
"r": 0.
|
206 |
-
"f": 0.
|
207 |
},
|
208 |
"preconj": {
|
209 |
-
"p": 0.
|
210 |
-
"r": 0.
|
211 |
-
"f": 0.
|
212 |
},
|
213 |
"oprd": {
|
214 |
-
"p": 0.
|
215 |
-
"r": 0.
|
216 |
-
"f": 0.
|
217 |
},
|
218 |
"parataxis": {
|
219 |
-
"p": 0.
|
220 |
-
"r": 0.
|
221 |
-
"f": 0.
|
222 |
},
|
223 |
"meta": {
|
224 |
-
"p": 0.
|
225 |
-
"r": 0.
|
226 |
-
"f": 0.
|
227 |
},
|
228 |
"csubjpass": {
|
229 |
-
"p": 0.
|
230 |
-
"r": 0.
|
231 |
-
"f": 0.
|
232 |
}
|
233 |
},
|
234 |
-
"ents_p": 0.
|
235 |
-
"ents_r": 0.
|
236 |
-
"ents_f": 0.
|
237 |
"ents_per_type": {
|
238 |
"DATE": {
|
239 |
-
"p": 0.
|
240 |
-
"r": 0.
|
241 |
-
"f": 0.
|
242 |
},
|
243 |
"GPE": {
|
244 |
-
"p": 0.
|
245 |
-
"r": 0.
|
246 |
-
"f": 0.
|
247 |
},
|
248 |
"ORDINAL": {
|
249 |
-
"p": 0.
|
250 |
-
"r": 0.
|
251 |
-
"f": 0.
|
252 |
-
},
|
253 |
-
"ORG": {
|
254 |
-
"p": 0.8032096817,
|
255 |
-
"r": 0.8093849417,
|
256 |
-
"f": 0.8062854879
|
257 |
},
|
258 |
"FAC": {
|
259 |
-
"p": 0.
|
260 |
-
"r": 0.
|
261 |
-
"f": 0.
|
|
|
|
|
|
|
|
|
|
|
262 |
},
|
263 |
"CARDINAL": {
|
264 |
-
"p": 0.
|
265 |
-
"r": 0.
|
266 |
-
"f": 0.
|
|
|
|
|
|
|
|
|
|
|
267 |
},
|
268 |
"PERSON": {
|
269 |
-
"p": 0.
|
270 |
-
"r": 0.
|
271 |
-
"f": 0.
|
272 |
},
|
273 |
"NORP": {
|
274 |
-
"p": 0.
|
275 |
-
"r": 0.
|
276 |
-
"f": 0.
|
277 |
-
},
|
278 |
-
"LOC": {
|
279 |
-
"p": 0.7210144928,
|
280 |
-
"r": 0.6337579618,
|
281 |
-
"f": 0.6745762712
|
282 |
},
|
283 |
"TIME": {
|
284 |
-
"p": 0.
|
285 |
-
"r": 0.
|
286 |
-
"f": 0.
|
287 |
},
|
288 |
"QUANTITY": {
|
289 |
-
"p": 0.
|
290 |
-
"r": 0.
|
291 |
-
"f": 0.
|
292 |
},
|
293 |
"EVENT": {
|
294 |
-
"p": 0.
|
295 |
-
"r": 0.
|
296 |
-
"f": 0.
|
297 |
},
|
298 |
"WORK_OF_ART": {
|
299 |
-
"p": 0.
|
300 |
-
"r": 0.
|
301 |
-
"f": 0.
|
302 |
-
},
|
303 |
-
"MONEY": {
|
304 |
-
"p": 0.9178743961,
|
305 |
-
"r": 0.8972845336,
|
306 |
-
"f": 0.9074626866
|
307 |
},
|
308 |
"LAW": {
|
309 |
-
"p": 0.
|
310 |
-
"r": 0.
|
311 |
-
"f": 0.
|
312 |
},
|
313 |
-
"
|
314 |
-
"p": 0.
|
315 |
-
"r": 0.
|
316 |
-
"f": 0.
|
317 |
},
|
318 |
-
"
|
319 |
-
"p": 0.
|
320 |
-
"r": 0.
|
321 |
-
"f": 0.
|
322 |
},
|
323 |
"LANGUAGE": {
|
324 |
-
"p": 0.
|
325 |
-
"r": 0.
|
326 |
-
"f": 0.
|
|
|
|
|
|
|
|
|
|
|
327 |
}
|
328 |
},
|
329 |
-
"speed":
|
330 |
}
|
|
|
1 |
{
|
2 |
+
"token_acc": 0.9993092439,
|
3 |
+
"token_p": 0.9956819193,
|
4 |
+
"token_r": 0.9957659295,
|
5 |
+
"token_f": 0.9957239226,
|
6 |
+
"tag_acc": 0.9726545475,
|
7 |
+
"sents_p": 0.9188657486,
|
8 |
+
"sents_r": 0.8935285969,
|
9 |
+
"sents_f": 0.9060200669,
|
10 |
+
"dep_uas": 0.9180803841,
|
11 |
+
"dep_las": 0.8996666011,
|
12 |
"dep_las_per_type": {
|
13 |
"prep": {
|
14 |
+
"p": 0.8545638793,
|
15 |
+
"r": 0.8639347372,
|
16 |
+
"f": 0.8592237589
|
17 |
},
|
18 |
"det": {
|
19 |
+
"p": 0.9767252604,
|
20 |
"r": 0.9787164642,
|
21 |
+
"f": 0.9777198485
|
22 |
},
|
23 |
"pobj": {
|
24 |
+
"p": 0.9624306803,
|
25 |
+
"r": 0.9677596701,
|
26 |
+
"f": 0.9650878189
|
27 |
},
|
28 |
"nsubj": {
|
29 |
+
"p": 0.9570163789,
|
30 |
+
"r": 0.9471631982,
|
31 |
+
"f": 0.9520642959
|
32 |
},
|
33 |
"aux": {
|
34 |
+
"p": 0.9800017776,
|
35 |
+
"r": 0.9815721535,
|
36 |
+
"f": 0.980786337
|
37 |
},
|
38 |
"advmod": {
|
39 |
+
"p": 0.8575166709,
|
40 |
+
"r": 0.8547030119,
|
41 |
+
"f": 0.8561075296
|
42 |
},
|
43 |
"relcl": {
|
44 |
+
"p": 0.7630824373,
|
45 |
+
"r": 0.7724963716,
|
46 |
+
"f": 0.7677605481
|
47 |
},
|
48 |
"root": {
|
49 |
+
"p": 0.9153887909,
|
50 |
+
"r": 0.8899663566,
|
51 |
+
"f": 0.9024985785
|
52 |
},
|
53 |
"xcomp": {
|
54 |
+
"p": 0.8755679832,
|
55 |
+
"r": 0.8991385499,
|
56 |
+
"f": 0.8871967416
|
57 |
},
|
58 |
"amod": {
|
59 |
+
"p": 0.9161676647,
|
60 |
+
"r": 0.9119533528,
|
61 |
+
"f": 0.9140556512
|
62 |
},
|
63 |
"compound": {
|
64 |
+
"p": 0.9153947513,
|
65 |
+
"r": 0.9285475607,
|
66 |
+
"f": 0.9219242466
|
67 |
},
|
68 |
"poss": {
|
69 |
+
"p": 0.9729241877,
|
70 |
+
"r": 0.9764492754,
|
71 |
+
"f": 0.9746835443
|
72 |
},
|
73 |
"ccomp": {
|
74 |
+
"p": 0.7720643231,
|
75 |
+
"r": 0.8409368635,
|
76 |
+
"f": 0.8050302203
|
77 |
},
|
78 |
"attr": {
|
79 |
+
"p": 0.9093147312,
|
80 |
+
"r": 0.9318755257,
|
81 |
+
"f": 0.9204569055
|
82 |
},
|
83 |
"case": {
|
84 |
+
"p": 0.977799704,
|
85 |
+
"r": 0.991991992,
|
86 |
+
"f": 0.9848447205
|
87 |
},
|
88 |
"mark": {
|
89 |
+
"p": 0.9035921817,
|
90 |
+
"r": 0.9064652888,
|
91 |
+
"f": 0.905026455
|
92 |
},
|
93 |
"intj": {
|
94 |
+
"p": 0.6711250983,
|
95 |
+
"r": 0.6249084249,
|
96 |
+
"f": 0.6471927162
|
97 |
},
|
98 |
"advcl": {
|
99 |
+
"p": 0.6756965944,
|
100 |
+
"r": 0.6595316041,
|
101 |
+
"f": 0.6675162482
|
102 |
},
|
103 |
"cc": {
|
104 |
+
"p": 0.8375451264,
|
105 |
+
"r": 0.8324363114,
|
106 |
+
"f": 0.8349829044
|
107 |
},
|
108 |
"neg": {
|
109 |
+
"p": 0.9480778832,
|
110 |
+
"r": 0.9528349222,
|
111 |
+
"f": 0.9504504505
|
112 |
},
|
113 |
"conj": {
|
114 |
+
"p": 0.7654719087,
|
115 |
+
"r": 0.77693857,
|
116 |
+
"f": 0.7711626164
|
117 |
},
|
118 |
"nsubjpass": {
|
119 |
+
"p": 0.9168804515,
|
120 |
+
"r": 0.9164102564,
|
121 |
+
"f": 0.9166452937
|
122 |
},
|
123 |
"auxpass": {
|
124 |
+
"p": 0.9459821429,
|
125 |
+
"r": 0.9653758542,
|
126 |
+
"f": 0.9555806088
|
127 |
},
|
128 |
"dobj": {
|
129 |
+
"p": 0.9223308565,
|
130 |
+
"r": 0.9396764682,
|
131 |
+
"f": 0.9309228705
|
132 |
},
|
133 |
"nummod": {
|
134 |
+
"p": 0.9368956743,
|
135 |
+
"r": 0.9297979798,
|
136 |
+
"f": 0.9333333333
|
137 |
},
|
138 |
"npadvmod": {
|
139 |
+
"p": 0.7781178271,
|
140 |
+
"r": 0.7225577265,
|
141 |
+
"f": 0.7493092651
|
142 |
},
|
143 |
"prt": {
|
144 |
+
"p": 0.816091954,
|
145 |
+
"r": 0.8906810036,
|
146 |
+
"f": 0.851756641
|
147 |
},
|
148 |
"pcomp": {
|
149 |
+
"p": 0.8699300699,
|
150 |
+
"r": 0.8711484594,
|
151 |
+
"f": 0.8705388383
|
152 |
},
|
153 |
"expl": {
|
154 |
+
"p": 0.983014862,
|
155 |
+
"r": 0.9914346895,
|
156 |
+
"f": 0.987206823
|
157 |
},
|
158 |
"acl": {
|
159 |
+
"p": 0.7332949309,
|
160 |
+
"r": 0.6944899073,
|
161 |
+
"f": 0.7133650883
|
162 |
},
|
163 |
"agent": {
|
164 |
+
"p": 0.8885135135,
|
165 |
+
"r": 0.9426523297,
|
166 |
+
"f": 0.9147826087
|
167 |
},
|
168 |
"dative": {
|
169 |
+
"p": 0.7847769029,
|
170 |
+
"r": 0.6857798165,
|
171 |
+
"f": 0.7319461444
|
172 |
},
|
173 |
"acomp": {
|
174 |
+
"p": 0.9046746104,
|
175 |
+
"r": 0.8952380952,
|
176 |
+
"f": 0.8999316161
|
177 |
},
|
178 |
"dep": {
|
179 |
+
"p": 0.4151624549,
|
180 |
+
"r": 0.1866883117,
|
181 |
+
"f": 0.2575587906
|
182 |
},
|
183 |
"csubj": {
|
184 |
+
"p": 0.6476683938,
|
185 |
+
"r": 0.7396449704,
|
186 |
+
"f": 0.6906077348
|
187 |
},
|
188 |
"quantmod": {
|
189 |
+
"p": 0.8682310469,
|
190 |
+
"r": 0.7814784728,
|
191 |
+
"f": 0.8225737495
|
192 |
},
|
193 |
"nmod": {
|
194 |
+
"p": 0.741078208,
|
195 |
+
"r": 0.5947592931,
|
196 |
+
"f": 0.6599053414
|
197 |
},
|
198 |
"appos": {
|
199 |
+
"p": 0.7215189873,
|
200 |
+
"r": 0.6676789588,
|
201 |
+
"f": 0.6935556557
|
202 |
},
|
203 |
"predet": {
|
204 |
+
"p": 0.8395061728,
|
205 |
+
"r": 0.8755364807,
|
206 |
+
"f": 0.8571428571
|
207 |
},
|
208 |
"preconj": {
|
209 |
+
"p": 0.5544554455,
|
210 |
+
"r": 0.6511627907,
|
211 |
+
"f": 0.5989304813
|
212 |
},
|
213 |
"oprd": {
|
214 |
+
"p": 0.8205980066,
|
215 |
+
"r": 0.7373134328,
|
216 |
+
"f": 0.7767295597
|
217 |
},
|
218 |
"parataxis": {
|
219 |
+
"p": 0.6121883657,
|
220 |
+
"r": 0.4793926247,
|
221 |
+
"f": 0.5377128954
|
222 |
},
|
223 |
"meta": {
|
224 |
+
"p": 0.7407407407,
|
225 |
+
"r": 0.3846153846,
|
226 |
+
"f": 0.5063291139
|
227 |
},
|
228 |
"csubjpass": {
|
229 |
+
"p": 0.7142857143,
|
230 |
+
"r": 0.8333333333,
|
231 |
+
"f": 0.7692307692
|
232 |
}
|
233 |
},
|
234 |
+
"ents_p": 0.8508041869,
|
235 |
+
"ents_r": 0.8344851763,
|
236 |
+
"ents_f": 0.8425656714,
|
237 |
"ents_per_type": {
|
238 |
"DATE": {
|
239 |
+
"p": 0.8732394366,
|
240 |
+
"r": 0.866031746,
|
241 |
+
"f": 0.8696206567
|
242 |
},
|
243 |
"GPE": {
|
244 |
+
"p": 0.9154443486,
|
245 |
+
"r": 0.8878661088,
|
246 |
+
"f": 0.90144435
|
247 |
},
|
248 |
"ORDINAL": {
|
249 |
+
"p": 0.7927927928,
|
250 |
+
"r": 0.8198757764,
|
251 |
+
"f": 0.8061068702
|
|
|
|
|
|
|
|
|
|
|
252 |
},
|
253 |
"FAC": {
|
254 |
+
"p": 0.4049586777,
|
255 |
+
"r": 0.3769230769,
|
256 |
+
"f": 0.390438247
|
257 |
+
},
|
258 |
+
"ORG": {
|
259 |
+
"p": 0.8038601982,
|
260 |
+
"r": 0.8170731707,
|
261 |
+
"f": 0.810412832
|
262 |
},
|
263 |
"CARDINAL": {
|
264 |
+
"p": 0.8222477064,
|
265 |
+
"r": 0.8525564804,
|
266 |
+
"f": 0.8371278459
|
267 |
+
},
|
268 |
+
"LOC": {
|
269 |
+
"p": 0.714801444,
|
270 |
+
"r": 0.6305732484,
|
271 |
+
"f": 0.6700507614
|
272 |
},
|
273 |
"PERSON": {
|
274 |
+
"p": 0.8572793883,
|
275 |
+
"r": 0.8782637076,
|
276 |
+
"f": 0.8676446881
|
277 |
},
|
278 |
"NORP": {
|
279 |
+
"p": 0.918652424,
|
280 |
+
"r": 0.8944,
|
281 |
+
"f": 0.9063640049
|
|
|
|
|
|
|
|
|
|
|
282 |
},
|
283 |
"TIME": {
|
284 |
+
"p": 0.7436708861,
|
285 |
+
"r": 0.6871345029,
|
286 |
+
"f": 0.7142857143
|
287 |
},
|
288 |
"QUANTITY": {
|
289 |
+
"p": 0.8308823529,
|
290 |
+
"r": 0.6208791209,
|
291 |
+
"f": 0.7106918239
|
292 |
},
|
293 |
"EVENT": {
|
294 |
+
"p": 0.5533980583,
|
295 |
+
"r": 0.3275862069,
|
296 |
+
"f": 0.4115523466
|
297 |
},
|
298 |
"WORK_OF_ART": {
|
299 |
+
"p": 0.4926470588,
|
300 |
+
"r": 0.3453608247,
|
301 |
+
"f": 0.4060606061
|
|
|
|
|
|
|
|
|
|
|
302 |
},
|
303 |
"LAW": {
|
304 |
+
"p": 0.58,
|
305 |
+
"r": 0.453125,
|
306 |
+
"f": 0.5087719298
|
307 |
},
|
308 |
+
"MONEY": {
|
309 |
+
"p": 0.9198564593,
|
310 |
+
"r": 0.9079102715,
|
311 |
+
"f": 0.9138443256
|
312 |
},
|
313 |
+
"PERCENT": {
|
314 |
+
"p": 0.9153354633,
|
315 |
+
"r": 0.8774885145,
|
316 |
+
"f": 0.8960125098
|
317 |
},
|
318 |
"LANGUAGE": {
|
319 |
+
"p": 0.7857142857,
|
320 |
+
"r": 0.6875,
|
321 |
+
"f": 0.7333333333
|
322 |
+
},
|
323 |
+
"PRODUCT": {
|
324 |
+
"p": 0.5795454545,
|
325 |
+
"r": 0.2417061611,
|
326 |
+
"f": 0.3411371237
|
327 |
}
|
328 |
},
|
329 |
+
"speed": 9738.3022066337
|
330 |
}
|
attribute_ruler/patterns
CHANGED
Binary files a/attribute_ruler/patterns and b/attribute_ruler/patterns differ
|
|
config.cfg
CHANGED
@@ -55,7 +55,7 @@ nO = null
|
|
55 |
@architectures = "spacy.MultiHashEmbed.v2"
|
56 |
width = 96
|
57 |
attrs = ["NORM","PREFIX","SUFFIX","SHAPE","SPACY"]
|
58 |
-
rows = [5000,
|
59 |
include_static_vectors = false
|
60 |
|
61 |
[components.ner.model.tok2vec.encode]
|
@@ -93,8 +93,9 @@ overwrite = false
|
|
93 |
scorer = {"@scorers":"spacy.senter_scorer.v1"}
|
94 |
|
95 |
[components.senter.model]
|
96 |
-
@architectures = "spacy.Tagger.
|
97 |
nO = null
|
|
|
98 |
|
99 |
[components.senter.model.tok2vec]
|
100 |
@architectures = "spacy.Tok2Vec.v2"
|
@@ -115,12 +116,14 @@ maxout_pieces = 2
|
|
115 |
|
116 |
[components.tagger]
|
117 |
factory = "tagger"
|
|
|
118 |
overwrite = false
|
119 |
scorer = {"@scorers":"spacy.tagger_scorer.v1"}
|
120 |
|
121 |
[components.tagger.model]
|
122 |
-
@architectures = "spacy.Tagger.
|
123 |
nO = null
|
|
|
124 |
|
125 |
[components.tagger.model.tok2vec]
|
126 |
@architectures = "spacy.Tok2VecListener.v1"
|
@@ -137,7 +140,7 @@ factory = "tok2vec"
|
|
137 |
@architectures = "spacy.MultiHashEmbed.v2"
|
138 |
width = ${components.tok2vec.model.encode:width}
|
139 |
attrs = ["NORM","PREFIX","SUFFIX","SHAPE","SPACY"]
|
140 |
-
rows = [5000,
|
141 |
include_static_vectors = false
|
142 |
|
143 |
[components.tok2vec.model.encode]
|
@@ -174,7 +177,7 @@ dropout = 0.1
|
|
174 |
accumulate_gradient = 1
|
175 |
patience = 5000
|
176 |
max_epochs = 0
|
177 |
-
max_steps =
|
178 |
eval_frequency = 1000
|
179 |
frozen_components = []
|
180 |
before_to_disk = null
|
|
|
55 |
@architectures = "spacy.MultiHashEmbed.v2"
|
56 |
width = 96
|
57 |
attrs = ["NORM","PREFIX","SUFFIX","SHAPE","SPACY"]
|
58 |
+
rows = [5000,1000,2500,2500,50]
|
59 |
include_static_vectors = false
|
60 |
|
61 |
[components.ner.model.tok2vec.encode]
|
|
|
93 |
scorer = {"@scorers":"spacy.senter_scorer.v1"}
|
94 |
|
95 |
[components.senter.model]
|
96 |
+
@architectures = "spacy.Tagger.v2"
|
97 |
nO = null
|
98 |
+
normalize = false
|
99 |
|
100 |
[components.senter.model.tok2vec]
|
101 |
@architectures = "spacy.Tok2Vec.v2"
|
|
|
116 |
|
117 |
[components.tagger]
|
118 |
factory = "tagger"
|
119 |
+
neg_prefix = "!"
|
120 |
overwrite = false
|
121 |
scorer = {"@scorers":"spacy.tagger_scorer.v1"}
|
122 |
|
123 |
[components.tagger.model]
|
124 |
+
@architectures = "spacy.Tagger.v2"
|
125 |
nO = null
|
126 |
+
normalize = false
|
127 |
|
128 |
[components.tagger.model.tok2vec]
|
129 |
@architectures = "spacy.Tok2VecListener.v1"
|
|
|
140 |
@architectures = "spacy.MultiHashEmbed.v2"
|
141 |
width = ${components.tok2vec.model.encode:width}
|
142 |
attrs = ["NORM","PREFIX","SUFFIX","SHAPE","SPACY"]
|
143 |
+
rows = [5000,1000,2500,2500,50]
|
144 |
include_static_vectors = false
|
145 |
|
146 |
[components.tok2vec.model.encode]
|
|
|
177 |
accumulate_gradient = 1
|
178 |
patience = 5000
|
179 |
max_epochs = 0
|
180 |
+
max_steps = 100000
|
181 |
eval_frequency = 1000
|
182 |
frozen_components = []
|
183 |
before_to_disk = null
|
en_core_web_sm-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 sha256:84d7d8059bfbf53c09b39139782f76cd6ac7064851e7799dcc685c06ebf5fd4f
|
3 |
+
size 12799907
|
meta.json
CHANGED
@@ -1,14 +1,14 @@
|
|
1 |
{
|
2 |
"lang":"en",
|
3 |
"name":"core_web_sm",
|
4 |
-
"version":"3.
|
5 |
"description":"English pipeline optimized for CPU. Components: tok2vec, tagger, parser, senter, ner, attribute_ruler, lemmatizer.",
|
6 |
"author":"Explosion",
|
7 |
"email":"contact@explosion.ai",
|
8 |
"url":"https://explosion.ai",
|
9 |
"license":"MIT",
|
10 |
-
"spacy_version":">=3.
|
11 |
-
"spacy_git_version":"
|
12 |
"vectors":{
|
13 |
"width":0,
|
14 |
"vectors":0,
|
@@ -117,10 +117,6 @@
|
|
117 |
"relcl",
|
118 |
"xcomp"
|
119 |
],
|
120 |
-
"senter":[
|
121 |
-
"I",
|
122 |
-
"S"
|
123 |
-
],
|
124 |
"attribute_ruler":[
|
125 |
|
126 |
],
|
@@ -169,334 +165,334 @@
|
|
169 |
"senter"
|
170 |
],
|
171 |
"performance":{
|
172 |
-
"token_acc":0.
|
173 |
-
"token_p":0.
|
174 |
-
"token_r":0.
|
175 |
-
"token_f":0.
|
176 |
-
"tag_acc":0.
|
177 |
-
"sents_p":0.
|
178 |
-
"sents_r":0.
|
179 |
-
"sents_f":0.
|
180 |
-
"dep_uas":0.
|
181 |
-
"dep_las":0.
|
182 |
"dep_las_per_type":{
|
183 |
"prep":{
|
184 |
-
"p":0.
|
185 |
-
"r":0.
|
186 |
-
"f":0.
|
187 |
},
|
188 |
"det":{
|
189 |
-
"p":0.
|
190 |
"r":0.9787164642,
|
191 |
-
"f":0.
|
192 |
},
|
193 |
"pobj":{
|
194 |
-
"p":0.
|
195 |
-
"r":0.
|
196 |
-
"f":0.
|
197 |
},
|
198 |
"nsubj":{
|
199 |
-
"p":0.
|
200 |
-
"r":0.
|
201 |
-
"f":0.
|
202 |
},
|
203 |
"aux":{
|
204 |
-
"p":0.
|
205 |
-
"r":0.
|
206 |
-
"f":0.
|
207 |
},
|
208 |
"advmod":{
|
209 |
-
"p":0.
|
210 |
-
"r":0.
|
211 |
-
"f":0.
|
212 |
},
|
213 |
"relcl":{
|
214 |
-
"p":0.
|
215 |
-
"r":0.
|
216 |
-
"f":0.
|
217 |
},
|
218 |
"root":{
|
219 |
-
"p":0.
|
220 |
-
"r":0.
|
221 |
-
"f":0.
|
222 |
},
|
223 |
"xcomp":{
|
224 |
-
"p":0.
|
225 |
-
"r":0.
|
226 |
-
"f":0.
|
227 |
},
|
228 |
"amod":{
|
229 |
-
"p":0.
|
230 |
-
"r":0.
|
231 |
-
"f":0.
|
232 |
},
|
233 |
"compound":{
|
234 |
-
"p":0.
|
235 |
-
"r":0.
|
236 |
-
"f":0.
|
237 |
},
|
238 |
"poss":{
|
239 |
-
"p":0.
|
240 |
-
"r":0.
|
241 |
-
"f":0.
|
242 |
},
|
243 |
"ccomp":{
|
244 |
-
"p":0.
|
245 |
-
"r":0.
|
246 |
-
"f":0.
|
247 |
},
|
248 |
"attr":{
|
249 |
-
"p":0.
|
250 |
-
"r":0.
|
251 |
-
"f":0.
|
252 |
},
|
253 |
"case":{
|
254 |
-
"p":0.
|
255 |
-
"r":0.
|
256 |
-
"f":0.
|
257 |
},
|
258 |
"mark":{
|
259 |
-
"p":0.
|
260 |
-
"r":0.
|
261 |
-
"f":0.
|
262 |
},
|
263 |
"intj":{
|
264 |
-
"p":0.
|
265 |
-
"r":0.
|
266 |
-
"f":0.
|
267 |
},
|
268 |
"advcl":{
|
269 |
-
"p":0.
|
270 |
-
"r":0.
|
271 |
-
"f":0.
|
272 |
},
|
273 |
"cc":{
|
274 |
-
"p":0.
|
275 |
-
"r":0.
|
276 |
-
"f":0.
|
277 |
},
|
278 |
"neg":{
|
279 |
-
"p":0.
|
280 |
-
"r":0.
|
281 |
-
"f":0.
|
282 |
},
|
283 |
"conj":{
|
284 |
-
"p":0.
|
285 |
-
"r":0.
|
286 |
-
"f":0.
|
287 |
},
|
288 |
"nsubjpass":{
|
289 |
-
"p":0.
|
290 |
-
"r":0.
|
291 |
-
"f":0.
|
292 |
},
|
293 |
"auxpass":{
|
294 |
-
"p":0.
|
295 |
-
"r":0.
|
296 |
-
"f":0.
|
297 |
},
|
298 |
"dobj":{
|
299 |
-
"p":0.
|
300 |
-
"r":0.
|
301 |
-
"f":0.
|
302 |
},
|
303 |
"nummod":{
|
304 |
-
"p":0.
|
305 |
-
"r":0.
|
306 |
-
"f":0.
|
307 |
},
|
308 |
"npadvmod":{
|
309 |
-
"p":0.
|
310 |
-
"r":0.
|
311 |
-
"f":0.
|
312 |
},
|
313 |
"prt":{
|
314 |
-
"p":0.
|
315 |
-
"r":0.
|
316 |
-
"f":0.
|
317 |
},
|
318 |
"pcomp":{
|
319 |
-
"p":0.
|
320 |
-
"r":0.
|
321 |
-
"f":0.
|
322 |
},
|
323 |
"expl":{
|
324 |
-
"p":0.
|
325 |
-
"r":0.
|
326 |
-
"f":0.
|
327 |
},
|
328 |
"acl":{
|
329 |
-
"p":0.
|
330 |
-
"r":0.
|
331 |
-
"f":0.
|
332 |
},
|
333 |
"agent":{
|
334 |
-
"p":0.
|
335 |
-
"r":0.
|
336 |
-
"f":0.
|
337 |
},
|
338 |
"dative":{
|
339 |
-
"p":0.
|
340 |
-
"r":0.
|
341 |
-
"f":0.
|
342 |
},
|
343 |
"acomp":{
|
344 |
-
"p":0.
|
345 |
-
"r":0.
|
346 |
-
"f":0.
|
347 |
},
|
348 |
"dep":{
|
349 |
-
"p":0.
|
350 |
-
"r":0.
|
351 |
-
"f":0.
|
352 |
},
|
353 |
"csubj":{
|
354 |
-
"p":0.
|
355 |
-
"r":0.
|
356 |
-
"f":0.
|
357 |
},
|
358 |
"quantmod":{
|
359 |
-
"p":0.
|
360 |
-
"r":0.
|
361 |
-
"f":0.
|
362 |
},
|
363 |
"nmod":{
|
364 |
-
"p":0.
|
365 |
-
"r":0.
|
366 |
-
"f":0.
|
367 |
},
|
368 |
"appos":{
|
369 |
-
"p":0.
|
370 |
-
"r":0.
|
371 |
-
"f":0.
|
372 |
},
|
373 |
"predet":{
|
374 |
-
"p":0.
|
375 |
-
"r":0.
|
376 |
-
"f":0.
|
377 |
},
|
378 |
"preconj":{
|
379 |
-
"p":0.
|
380 |
-
"r":0.
|
381 |
-
"f":0.
|
382 |
},
|
383 |
"oprd":{
|
384 |
-
"p":0.
|
385 |
-
"r":0.
|
386 |
-
"f":0.
|
387 |
},
|
388 |
"parataxis":{
|
389 |
-
"p":0.
|
390 |
-
"r":0.
|
391 |
-
"f":0.
|
392 |
},
|
393 |
"meta":{
|
394 |
-
"p":0.
|
395 |
-
"r":0.
|
396 |
-
"f":0.
|
397 |
},
|
398 |
"csubjpass":{
|
399 |
-
"p":0.
|
400 |
-
"r":0.
|
401 |
-
"f":0.
|
402 |
}
|
403 |
},
|
404 |
-
"ents_p":0.
|
405 |
-
"ents_r":0.
|
406 |
-
"ents_f":0.
|
407 |
"ents_per_type":{
|
408 |
"DATE":{
|
409 |
-
"p":0.
|
410 |
-
"r":0.
|
411 |
-
"f":0.
|
412 |
},
|
413 |
"GPE":{
|
414 |
-
"p":0.
|
415 |
-
"r":0.
|
416 |
-
"f":0.
|
417 |
},
|
418 |
"ORDINAL":{
|
419 |
-
"p":0.
|
420 |
-
"r":0.
|
421 |
-
"f":0.
|
422 |
-
},
|
423 |
-
"ORG":{
|
424 |
-
"p":0.8032096817,
|
425 |
-
"r":0.8093849417,
|
426 |
-
"f":0.8062854879
|
427 |
},
|
428 |
"FAC":{
|
429 |
-
"p":0.
|
430 |
-
"r":0.
|
431 |
-
"f":0.
|
|
|
|
|
|
|
|
|
|
|
432 |
},
|
433 |
"CARDINAL":{
|
434 |
-
"p":0.
|
435 |
-
"r":0.
|
436 |
-
"f":0.
|
|
|
|
|
|
|
|
|
|
|
437 |
},
|
438 |
"PERSON":{
|
439 |
-
"p":0.
|
440 |
-
"r":0.
|
441 |
-
"f":0.
|
442 |
},
|
443 |
"NORP":{
|
444 |
-
"p":0.
|
445 |
-
"r":0.
|
446 |
-
"f":0.
|
447 |
-
},
|
448 |
-
"LOC":{
|
449 |
-
"p":0.7210144928,
|
450 |
-
"r":0.6337579618,
|
451 |
-
"f":0.6745762712
|
452 |
},
|
453 |
"TIME":{
|
454 |
-
"p":0.
|
455 |
-
"r":0.
|
456 |
-
"f":0.
|
457 |
},
|
458 |
"QUANTITY":{
|
459 |
-
"p":0.
|
460 |
-
"r":0.
|
461 |
-
"f":0.
|
462 |
},
|
463 |
"EVENT":{
|
464 |
-
"p":0.
|
465 |
-
"r":0.
|
466 |
-
"f":0.
|
467 |
},
|
468 |
"WORK_OF_ART":{
|
469 |
-
"p":0.
|
470 |
-
"r":0.
|
471 |
-
"f":0.
|
472 |
-
},
|
473 |
-
"MONEY":{
|
474 |
-
"p":0.9178743961,
|
475 |
-
"r":0.8972845336,
|
476 |
-
"f":0.9074626866
|
477 |
},
|
478 |
"LAW":{
|
479 |
-
"p":0.
|
480 |
-
"r":0.
|
481 |
-
"f":0.
|
482 |
},
|
483 |
-
"
|
484 |
-
"p":0.
|
485 |
-
"r":0.
|
486 |
-
"f":0.
|
487 |
},
|
488 |
-
"
|
489 |
-
"p":0.
|
490 |
-
"r":0.
|
491 |
-
"f":0.
|
492 |
},
|
493 |
"LANGUAGE":{
|
494 |
-
"p":0.
|
495 |
-
"r":0.
|
496 |
-
"f":0.
|
|
|
|
|
|
|
|
|
|
|
497 |
}
|
498 |
},
|
499 |
-
"speed":
|
500 |
},
|
501 |
"sources":[
|
502 |
{
|
|
|
1 |
{
|
2 |
"lang":"en",
|
3 |
"name":"core_web_sm",
|
4 |
+
"version":"3.3.0",
|
5 |
"description":"English pipeline optimized for CPU. Components: tok2vec, tagger, parser, senter, ner, attribute_ruler, lemmatizer.",
|
6 |
"author":"Explosion",
|
7 |
"email":"contact@explosion.ai",
|
8 |
"url":"https://explosion.ai",
|
9 |
"license":"MIT",
|
10 |
+
"spacy_version":">=3.3.0.dev0,<3.4.0",
|
11 |
+
"spacy_git_version":"849bef2de",
|
12 |
"vectors":{
|
13 |
"width":0,
|
14 |
"vectors":0,
|
|
|
117 |
"relcl",
|
118 |
"xcomp"
|
119 |
],
|
|
|
|
|
|
|
|
|
120 |
"attribute_ruler":[
|
121 |
|
122 |
],
|
|
|
165 |
"senter"
|
166 |
],
|
167 |
"performance":{
|
168 |
+
"token_acc":0.9993092439,
|
169 |
+
"token_p":0.9956819193,
|
170 |
+
"token_r":0.9957659295,
|
171 |
+
"token_f":0.9957239226,
|
172 |
+
"tag_acc":0.9726545475,
|
173 |
+
"sents_p":0.9188657486,
|
174 |
+
"sents_r":0.8935285969,
|
175 |
+
"sents_f":0.9060200669,
|
176 |
+
"dep_uas":0.9180803841,
|
177 |
+
"dep_las":0.8996666011,
|
178 |
"dep_las_per_type":{
|
179 |
"prep":{
|
180 |
+
"p":0.8545638793,
|
181 |
+
"r":0.8639347372,
|
182 |
+
"f":0.8592237589
|
183 |
},
|
184 |
"det":{
|
185 |
+
"p":0.9767252604,
|
186 |
"r":0.9787164642,
|
187 |
+
"f":0.9777198485
|
188 |
},
|
189 |
"pobj":{
|
190 |
+
"p":0.9624306803,
|
191 |
+
"r":0.9677596701,
|
192 |
+
"f":0.9650878189
|
193 |
},
|
194 |
"nsubj":{
|
195 |
+
"p":0.9570163789,
|
196 |
+
"r":0.9471631982,
|
197 |
+
"f":0.9520642959
|
198 |
},
|
199 |
"aux":{
|
200 |
+
"p":0.9800017776,
|
201 |
+
"r":0.9815721535,
|
202 |
+
"f":0.980786337
|
203 |
},
|
204 |
"advmod":{
|
205 |
+
"p":0.8575166709,
|
206 |
+
"r":0.8547030119,
|
207 |
+
"f":0.8561075296
|
208 |
},
|
209 |
"relcl":{
|
210 |
+
"p":0.7630824373,
|
211 |
+
"r":0.7724963716,
|
212 |
+
"f":0.7677605481
|
213 |
},
|
214 |
"root":{
|
215 |
+
"p":0.9153887909,
|
216 |
+
"r":0.8899663566,
|
217 |
+
"f":0.9024985785
|
218 |
},
|
219 |
"xcomp":{
|
220 |
+
"p":0.8755679832,
|
221 |
+
"r":0.8991385499,
|
222 |
+
"f":0.8871967416
|
223 |
},
|
224 |
"amod":{
|
225 |
+
"p":0.9161676647,
|
226 |
+
"r":0.9119533528,
|
227 |
+
"f":0.9140556512
|
228 |
},
|
229 |
"compound":{
|
230 |
+
"p":0.9153947513,
|
231 |
+
"r":0.9285475607,
|
232 |
+
"f":0.9219242466
|
233 |
},
|
234 |
"poss":{
|
235 |
+
"p":0.9729241877,
|
236 |
+
"r":0.9764492754,
|
237 |
+
"f":0.9746835443
|
238 |
},
|
239 |
"ccomp":{
|
240 |
+
"p":0.7720643231,
|
241 |
+
"r":0.8409368635,
|
242 |
+
"f":0.8050302203
|
243 |
},
|
244 |
"attr":{
|
245 |
+
"p":0.9093147312,
|
246 |
+
"r":0.9318755257,
|
247 |
+
"f":0.9204569055
|
248 |
},
|
249 |
"case":{
|
250 |
+
"p":0.977799704,
|
251 |
+
"r":0.991991992,
|
252 |
+
"f":0.9848447205
|
253 |
},
|
254 |
"mark":{
|
255 |
+
"p":0.9035921817,
|
256 |
+
"r":0.9064652888,
|
257 |
+
"f":0.905026455
|
258 |
},
|
259 |
"intj":{
|
260 |
+
"p":0.6711250983,
|
261 |
+
"r":0.6249084249,
|
262 |
+
"f":0.6471927162
|
263 |
},
|
264 |
"advcl":{
|
265 |
+
"p":0.6756965944,
|
266 |
+
"r":0.6595316041,
|
267 |
+
"f":0.6675162482
|
268 |
},
|
269 |
"cc":{
|
270 |
+
"p":0.8375451264,
|
271 |
+
"r":0.8324363114,
|
272 |
+
"f":0.8349829044
|
273 |
},
|
274 |
"neg":{
|
275 |
+
"p":0.9480778832,
|
276 |
+
"r":0.9528349222,
|
277 |
+
"f":0.9504504505
|
278 |
},
|
279 |
"conj":{
|
280 |
+
"p":0.7654719087,
|
281 |
+
"r":0.77693857,
|
282 |
+
"f":0.7711626164
|
283 |
},
|
284 |
"nsubjpass":{
|
285 |
+
"p":0.9168804515,
|
286 |
+
"r":0.9164102564,
|
287 |
+
"f":0.9166452937
|
288 |
},
|
289 |
"auxpass":{
|
290 |
+
"p":0.9459821429,
|
291 |
+
"r":0.9653758542,
|
292 |
+
"f":0.9555806088
|
293 |
},
|
294 |
"dobj":{
|
295 |
+
"p":0.9223308565,
|
296 |
+
"r":0.9396764682,
|
297 |
+
"f":0.9309228705
|
298 |
},
|
299 |
"nummod":{
|
300 |
+
"p":0.9368956743,
|
301 |
+
"r":0.9297979798,
|
302 |
+
"f":0.9333333333
|
303 |
},
|
304 |
"npadvmod":{
|
305 |
+
"p":0.7781178271,
|
306 |
+
"r":0.7225577265,
|
307 |
+
"f":0.7493092651
|
308 |
},
|
309 |
"prt":{
|
310 |
+
"p":0.816091954,
|
311 |
+
"r":0.8906810036,
|
312 |
+
"f":0.851756641
|
313 |
},
|
314 |
"pcomp":{
|
315 |
+
"p":0.8699300699,
|
316 |
+
"r":0.8711484594,
|
317 |
+
"f":0.8705388383
|
318 |
},
|
319 |
"expl":{
|
320 |
+
"p":0.983014862,
|
321 |
+
"r":0.9914346895,
|
322 |
+
"f":0.987206823
|
323 |
},
|
324 |
"acl":{
|
325 |
+
"p":0.7332949309,
|
326 |
+
"r":0.6944899073,
|
327 |
+
"f":0.7133650883
|
328 |
},
|
329 |
"agent":{
|
330 |
+
"p":0.8885135135,
|
331 |
+
"r":0.9426523297,
|
332 |
+
"f":0.9147826087
|
333 |
},
|
334 |
"dative":{
|
335 |
+
"p":0.7847769029,
|
336 |
+
"r":0.6857798165,
|
337 |
+
"f":0.7319461444
|
338 |
},
|
339 |
"acomp":{
|
340 |
+
"p":0.9046746104,
|
341 |
+
"r":0.8952380952,
|
342 |
+
"f":0.8999316161
|
343 |
},
|
344 |
"dep":{
|
345 |
+
"p":0.4151624549,
|
346 |
+
"r":0.1866883117,
|
347 |
+
"f":0.2575587906
|
348 |
},
|
349 |
"csubj":{
|
350 |
+
"p":0.6476683938,
|
351 |
+
"r":0.7396449704,
|
352 |
+
"f":0.6906077348
|
353 |
},
|
354 |
"quantmod":{
|
355 |
+
"p":0.8682310469,
|
356 |
+
"r":0.7814784728,
|
357 |
+
"f":0.8225737495
|
358 |
},
|
359 |
"nmod":{
|
360 |
+
"p":0.741078208,
|
361 |
+
"r":0.5947592931,
|
362 |
+
"f":0.6599053414
|
363 |
},
|
364 |
"appos":{
|
365 |
+
"p":0.7215189873,
|
366 |
+
"r":0.6676789588,
|
367 |
+
"f":0.6935556557
|
368 |
},
|
369 |
"predet":{
|
370 |
+
"p":0.8395061728,
|
371 |
+
"r":0.8755364807,
|
372 |
+
"f":0.8571428571
|
373 |
},
|
374 |
"preconj":{
|
375 |
+
"p":0.5544554455,
|
376 |
+
"r":0.6511627907,
|
377 |
+
"f":0.5989304813
|
378 |
},
|
379 |
"oprd":{
|
380 |
+
"p":0.8205980066,
|
381 |
+
"r":0.7373134328,
|
382 |
+
"f":0.7767295597
|
383 |
},
|
384 |
"parataxis":{
|
385 |
+
"p":0.6121883657,
|
386 |
+
"r":0.4793926247,
|
387 |
+
"f":0.5377128954
|
388 |
},
|
389 |
"meta":{
|
390 |
+
"p":0.7407407407,
|
391 |
+
"r":0.3846153846,
|
392 |
+
"f":0.5063291139
|
393 |
},
|
394 |
"csubjpass":{
|
395 |
+
"p":0.7142857143,
|
396 |
+
"r":0.8333333333,
|
397 |
+
"f":0.7692307692
|
398 |
}
|
399 |
},
|
400 |
+
"ents_p":0.8508041869,
|
401 |
+
"ents_r":0.8344851763,
|
402 |
+
"ents_f":0.8425656714,
|
403 |
"ents_per_type":{
|
404 |
"DATE":{
|
405 |
+
"p":0.8732394366,
|
406 |
+
"r":0.866031746,
|
407 |
+
"f":0.8696206567
|
408 |
},
|
409 |
"GPE":{
|
410 |
+
"p":0.9154443486,
|
411 |
+
"r":0.8878661088,
|
412 |
+
"f":0.90144435
|
413 |
},
|
414 |
"ORDINAL":{
|
415 |
+
"p":0.7927927928,
|
416 |
+
"r":0.8198757764,
|
417 |
+
"f":0.8061068702
|
|
|
|
|
|
|
|
|
|
|
418 |
},
|
419 |
"FAC":{
|
420 |
+
"p":0.4049586777,
|
421 |
+
"r":0.3769230769,
|
422 |
+
"f":0.390438247
|
423 |
+
},
|
424 |
+
"ORG":{
|
425 |
+
"p":0.8038601982,
|
426 |
+
"r":0.8170731707,
|
427 |
+
"f":0.810412832
|
428 |
},
|
429 |
"CARDINAL":{
|
430 |
+
"p":0.8222477064,
|
431 |
+
"r":0.8525564804,
|
432 |
+
"f":0.8371278459
|
433 |
+
},
|
434 |
+
"LOC":{
|
435 |
+
"p":0.714801444,
|
436 |
+
"r":0.6305732484,
|
437 |
+
"f":0.6700507614
|
438 |
},
|
439 |
"PERSON":{
|
440 |
+
"p":0.8572793883,
|
441 |
+
"r":0.8782637076,
|
442 |
+
"f":0.8676446881
|
443 |
},
|
444 |
"NORP":{
|
445 |
+
"p":0.918652424,
|
446 |
+
"r":0.8944,
|
447 |
+
"f":0.9063640049
|
|
|
|
|
|
|
|
|
|
|
448 |
},
|
449 |
"TIME":{
|
450 |
+
"p":0.7436708861,
|
451 |
+
"r":0.6871345029,
|
452 |
+
"f":0.7142857143
|
453 |
},
|
454 |
"QUANTITY":{
|
455 |
+
"p":0.8308823529,
|
456 |
+
"r":0.6208791209,
|
457 |
+
"f":0.7106918239
|
458 |
},
|
459 |
"EVENT":{
|
460 |
+
"p":0.5533980583,
|
461 |
+
"r":0.3275862069,
|
462 |
+
"f":0.4115523466
|
463 |
},
|
464 |
"WORK_OF_ART":{
|
465 |
+
"p":0.4926470588,
|
466 |
+
"r":0.3453608247,
|
467 |
+
"f":0.4060606061
|
|
|
|
|
|
|
|
|
|
|
468 |
},
|
469 |
"LAW":{
|
470 |
+
"p":0.58,
|
471 |
+
"r":0.453125,
|
472 |
+
"f":0.5087719298
|
473 |
},
|
474 |
+
"MONEY":{
|
475 |
+
"p":0.9198564593,
|
476 |
+
"r":0.9079102715,
|
477 |
+
"f":0.9138443256
|
478 |
},
|
479 |
+
"PERCENT":{
|
480 |
+
"p":0.9153354633,
|
481 |
+
"r":0.8774885145,
|
482 |
+
"f":0.8960125098
|
483 |
},
|
484 |
"LANGUAGE":{
|
485 |
+
"p":0.7857142857,
|
486 |
+
"r":0.6875,
|
487 |
+
"f":0.7333333333
|
488 |
+
},
|
489 |
+
"PRODUCT":{
|
490 |
+
"p":0.5795454545,
|
491 |
+
"r":0.2417061611,
|
492 |
+
"f":0.3411371237
|
493 |
}
|
494 |
},
|
495 |
+
"speed":9738.3022066337
|
496 |
},
|
497 |
"sources":[
|
498 |
{
|
ner/model
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:b150d51c7d759de938bcd86aa40744117371278ddae98c9d909939c38af0000f
|
3 |
+
size 6284763
|
parser/model
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 319909
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4dffb1da181e83f7742b1a1f8060b9cb039c2136f49df9d7a94e24a4529b428f
|
3 |
size 319909
|
parser/moves
CHANGED
@@ -1,2 +1,2 @@
|
|
1 |
��moves�
|
2 |
-
{"0":{"":
|
|
|
1 |
��moves�
|
2 |
+
{"0":{"":994267},"1":{"":990803},"2":{"det":172595,"nsubj":165748,"compound":116623,"amod":105184,"aux":86667,"punct":65478,"advmod":62763,"poss":36443,"mark":27941,"nummod":22598,"auxpass":15594,"prep":14001,"nsubjpass":13856,"neg":12357,"cc":10739,"nmod":9562,"advcl":9062,"npadvmod":8168,"quantmod":7101,"intj":6464,"ccomp":5896,"dobj":3427,"expl":3360,"dep":2806,"predet":1944,"parataxis":1837,"csubj":1428,"preconj":621,"pobj||prep":616,"attr":578,"meta":376,"advmod||conj":368,"dobj||xcomp":352,"acomp":284,"nsubj||ccomp":224,"dative":206,"advmod||xcomp":149,"dobj||ccomp":70,"csubjpass":64,"dobj||conj":62,"prep||conj":51,"acl":48,"prep||nsubj":41,"prep||dobj":36,"xcomp":34,"advmod||ccomp":32,"oprd":31},"3":{"punct":183790,"pobj":182191,"prep":174008,"dobj":89615,"conj":59687,"cc":51930,"ccomp":30385,"advmod":22861,"xcomp":21021,"relcl":20969,"advcl":19828,"attr":17741,"acomp":16922,"appos":15265,"case":13388,"acl":12085,"pcomp":10324,"npadvmod":9796,"prt":8179,"agent":3903,"dative":3866,"nsubj":3470,"neg":2906,"amod":2839,"intj":2819,"nummod":2732,"oprd":2301,"dep":1487,"parataxis":1261,"quantmod":319,"nmod":294,"acl||dobj":200,"prep||dobj":190,"prep||nsubj":162,"acl||nsubj":159,"appos||nsubj":145,"relcl||dobj":134,"relcl||nsubj":111,"aux":103,"expl":96,"meta":92,"appos||dobj":86,"preconj":71,"csubj":65,"prep||nsubjpass":55,"prep||advmod":54,"prep||acomp":53,"det":51,"nsubjpass":45,"relcl||pobj":42,"acl||nsubjpass":42,"mark":40,"auxpass":39,"prep||pobj":36,"relcl||nsubjpass":32,"appos||nsubjpass":31},"4":{"ROOT":111664}}�cfg��neg_key�
|
senter/model
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:f0dcd17fe48150d6202c6e499a296f713f29506ffb0fdc4fe76631b9c84df0a6
|
3 |
+
size 197089
|
tagger/cfg
CHANGED
@@ -50,5 +50,6 @@
|
|
50 |
"XX",
|
51 |
"``"
|
52 |
],
|
|
|
53 |
"overwrite":false
|
54 |
}
|
|
|
50 |
"XX",
|
51 |
"``"
|
52 |
],
|
53 |
+
"neg_prefix":"!",
|
54 |
"overwrite":false
|
55 |
}
|
tagger/model
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:a8decebdc61f212d8373cbd9255b6593461c29db529daa88ca1b222fb576f8a3
|
3 |
+
size 19441
|
tok2vec/model
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 sha256:1948c4cf0d3b8bfc58386f9f62a3cfdbffc3ff29e13d960a5392896e06d60a2f
|
3 |
+
size 6139229
|
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 sha256:9fbd4af89668f1aa5e362c98949666f92186f169e0941afc4f00172d4f25a2cd
|
3 |
+
size 1089443
|