etikaj-digital
commited on
Commit
•
5af7e49
1
Parent(s):
6e59214
Update spaCy pipeline
Browse files- config.cfg +1 -33
- en_statistics-any-py3-none-any.whl +2 -2
- meta.json +4 -26
config.cfg
CHANGED
@@ -12,7 +12,7 @@ seed = 0
|
|
12 |
|
13 |
[nlp]
|
14 |
lang = "en"
|
15 |
-
pipeline = ["tok2vec","tagger","parser","senter","attribute_ruler","lemmatizer","
|
16 |
disabled = ["senter"]
|
17 |
before_creation = null
|
18 |
after_creation = null
|
@@ -35,38 +35,6 @@ mode = "rule"
|
|
35 |
model = null
|
36 |
overwrite = false
|
37 |
|
38 |
-
[components.ner]
|
39 |
-
factory = "ner"
|
40 |
-
incorrect_spans_key = null
|
41 |
-
moves = null
|
42 |
-
update_with_oracle_cut_size = 100
|
43 |
-
|
44 |
-
[components.ner.model]
|
45 |
-
@architectures = "spacy.TransitionBasedParser.v2"
|
46 |
-
state_type = "ner"
|
47 |
-
extra_state_tokens = false
|
48 |
-
hidden_width = 64
|
49 |
-
maxout_pieces = 2
|
50 |
-
use_upper = true
|
51 |
-
nO = null
|
52 |
-
|
53 |
-
[components.ner.model.tok2vec]
|
54 |
-
@architectures = "spacy.Tok2Vec.v2"
|
55 |
-
|
56 |
-
[components.ner.model.tok2vec.embed]
|
57 |
-
@architectures = "spacy.MultiHashEmbed.v2"
|
58 |
-
width = 96
|
59 |
-
attrs = ["NORM","PREFIX","SUFFIX","SHAPE"]
|
60 |
-
rows = [5000,2500,2500,2500]
|
61 |
-
include_static_vectors = true
|
62 |
-
|
63 |
-
[components.ner.model.tok2vec.encode]
|
64 |
-
@architectures = "spacy.MaxoutWindowEncoder.v2"
|
65 |
-
width = 96
|
66 |
-
depth = 4
|
67 |
-
window_size = 1
|
68 |
-
maxout_pieces = 3
|
69 |
-
|
70 |
[components.parser]
|
71 |
factory = "parser"
|
72 |
learn_tokens = false
|
|
|
12 |
|
13 |
[nlp]
|
14 |
lang = "en"
|
15 |
+
pipeline = ["tok2vec","tagger","parser","senter","attribute_ruler","lemmatizer","syllables","formality","readability"]
|
16 |
disabled = ["senter"]
|
17 |
before_creation = null
|
18 |
after_creation = null
|
|
|
35 |
model = null
|
36 |
overwrite = false
|
37 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
38 |
[components.parser]
|
39 |
factory = "parser"
|
40 |
learn_tokens = false
|
en_statistics-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 sha256:994b15428304a8a21713c69e71231a922a4ae0b6344ab1e3088c81dd0d9f1258
|
3 |
+
size 45410416
|
meta.json
CHANGED
@@ -2,10 +2,10 @@
|
|
2 |
"lang":"en",
|
3 |
"name":"statistics",
|
4 |
"version":"0.0.1",
|
5 |
-
"description":"
|
6 |
-
"author":"
|
7 |
-
"email":"",
|
8 |
-
"url":"",
|
9 |
"license":"MIT",
|
10 |
"spacy_version":">=3.1.1,<3.2.0",
|
11 |
"spacy_git_version":"caba63b74",
|
@@ -126,26 +126,6 @@
|
|
126 |
],
|
127 |
"lemmatizer":[
|
128 |
|
129 |
-
],
|
130 |
-
"ner":[
|
131 |
-
"CARDINAL",
|
132 |
-
"DATE",
|
133 |
-
"EVENT",
|
134 |
-
"FAC",
|
135 |
-
"GPE",
|
136 |
-
"LANGUAGE",
|
137 |
-
"LAW",
|
138 |
-
"LOC",
|
139 |
-
"MONEY",
|
140 |
-
"NORP",
|
141 |
-
"ORDINAL",
|
142 |
-
"ORG",
|
143 |
-
"PERCENT",
|
144 |
-
"PERSON",
|
145 |
-
"PRODUCT",
|
146 |
-
"QUANTITY",
|
147 |
-
"TIME",
|
148 |
-
"WORK_OF_ART"
|
149 |
]
|
150 |
},
|
151 |
"pipeline":[
|
@@ -154,7 +134,6 @@
|
|
154 |
"parser",
|
155 |
"attribute_ruler",
|
156 |
"lemmatizer",
|
157 |
-
"ner",
|
158 |
"syllables",
|
159 |
"formality",
|
160 |
"readability"
|
@@ -166,7 +145,6 @@
|
|
166 |
"senter",
|
167 |
"attribute_ruler",
|
168 |
"lemmatizer",
|
169 |
-
"ner",
|
170 |
"syllables",
|
171 |
"formality",
|
172 |
"readability"
|
|
|
2 |
"lang":"en",
|
3 |
"name":"statistics",
|
4 |
"version":"0.0.1",
|
5 |
+
"description":"Calculate statistics including readability and formality.",
|
6 |
+
"author":"Chris Knowles",
|
7 |
+
"email":"knowles.chris.d@gmail.com",
|
8 |
+
"url":"https://explosion.ai",
|
9 |
"license":"MIT",
|
10 |
"spacy_version":">=3.1.1,<3.2.0",
|
11 |
"spacy_git_version":"caba63b74",
|
|
|
126 |
],
|
127 |
"lemmatizer":[
|
128 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
129 |
]
|
130 |
},
|
131 |
"pipeline":[
|
|
|
134 |
"parser",
|
135 |
"attribute_ruler",
|
136 |
"lemmatizer",
|
|
|
137 |
"syllables",
|
138 |
"formality",
|
139 |
"readability"
|
|
|
145 |
"senter",
|
146 |
"attribute_ruler",
|
147 |
"lemmatizer",
|
|
|
148 |
"syllables",
|
149 |
"formality",
|
150 |
"readability"
|