nb_core_news_md / meta.json
adrianeboyd's picture
Update spaCy pipeline
940f852
raw
history blame
19.4 kB
{
"lang":"nb",
"name":"core_news_md",
"version":"3.5.0",
"description":"Norwegian (Bokm\u00e5l) pipeline optimized for CPU. Components: tok2vec, morphologizer, parser, lemmatizer (trainable_lemmatizer), senter, ner, attribute_ruler.",
"author":"Explosion",
"email":"contact@explosion.ai",
"url":"https://explosion.ai",
"license":"MIT",
"spacy_version":">=3.5.0,<3.6.0",
"spacy_git_version":"9e0322de1",
"vectors":{
"width":300,
"vectors":20000,
"keys":500000,
"name":"nb_vectors"
},
"labels":{
"tok2vec":[
],
"morphologizer":[
"Definite=Ind|Gender=Neut|Number=Sing|POS=NOUN",
"POS=CCONJ",
"Definite=Ind|Gender=Masc|Number=Sing|POS=NOUN",
"POS=SCONJ",
"Definite=Def|Gender=Masc|Number=Sing|POS=NOUN",
"Definite=Ind|Gender=Neut|Number=Plur|POS=NOUN",
"POS=PUNCT",
"Mood=Ind|POS=VERB|Tense=Past|VerbForm=Fin",
"POS=ADP",
"Gender=Masc|Number=Sing|POS=DET|PronType=Dem",
"Definite=Def|Degree=Pos|Number=Sing|POS=ADJ",
"POS=PROPN",
"POS=X",
"Mood=Ind|POS=VERB|Tense=Pres|VerbForm=Fin",
"Definite=Def|Gender=Neut|Number=Sing|POS=NOUN",
"POS=PRON|PronType=Rel",
"Mood=Ind|POS=AUX|Tense=Pres|VerbForm=Fin",
"Definite=Ind|Gender=Neut|Number=Sing|POS=ADJ|VerbForm=Part",
"Definite=Ind|Degree=Pos|Number=Sing|POS=ADJ",
"Definite=Ind|Gender=Fem|Number=Sing|POS=NOUN",
"Number=Plur|POS=ADJ|VerbForm=Part",
"Definite=Ind|Gender=Fem|Number=Plur|POS=NOUN",
"POS=ADV",
"Gender=Neut|Number=Sing|POS=PRON|Person=3|PronType=Prs",
"Definite=Ind|Number=Sing|POS=ADJ|VerbForm=Part",
"POS=VERB|VerbForm=Part",
"Definite=Ind|Gender=Masc|Number=Plur|POS=NOUN",
"Definite=Ind|Degree=Pos|Gender=Neut|Number=Sing|POS=ADJ",
"Degree=Pos|Number=Plur|POS=ADJ",
"NumType=Card|Number=Plur|POS=NUM",
"Definite=Def|Gender=Masc|Number=Plur|POS=NOUN",
"Case=Acc|POS=PRON|PronType=Prs|Reflex=Yes",
"Case=Gen|Definite=Ind|Gender=Neut|Number=Sing|POS=NOUN",
"POS=PART",
"POS=VERB|VerbForm=Inf",
"Case=Nom|Number=Plur|POS=PRON|Person=3|PronType=Prs",
"Mood=Ind|POS=AUX|Tense=Past|VerbForm=Fin",
"Gender=Fem|POS=PROPN",
"POS=NOUN",
"Gender=Masc|POS=PROPN",
"Gender=Neut|Number=Sing|POS=DET|PronType=Dem",
"Gender=Masc|Number=Sing|POS=DET|PronType=Art",
"Case=Gen|Definite=Def|Gender=Masc|Number=Sing|POS=NOUN",
"Abbr=Yes|POS=PROPN",
"POS=PART|Polarity=Neg",
"Number=Plur|POS=PRON|Poss=Yes|PronType=Prs",
"Case=Gen|Definite=Ind|Gender=Neut|Number=Plur|POS=NOUN",
"Case=Gen|POS=PROPN",
"Gender=Fem|Number=Sing|POS=DET|PronType=Dem",
"Gender=Masc|Number=Sing|POS=PRON|Poss=Yes|PronType=Prs",
"Definite=Def|Degree=Sup|POS=ADJ",
"Case=Gen|Gender=Fem|POS=PROPN",
"Number=Plur|POS=DET|PronType=Dem",
"Case=Gen|Definite=Def|Gender=Neut|Number=Sing|POS=NOUN",
"Definite=Ind|Degree=Sup|POS=ADJ",
"Definite=Def|Gender=Fem|Number=Plur|POS=NOUN",
"Gender=Neut|POS=PROPN",
"Number=Plur|POS=DET|PronType=Int",
"Definite=Def|Gender=Neut|Number=Plur|POS=NOUN",
"Definite=Def|POS=DET|PronType=Dem",
"Gender=Neut|Number=Sing|POS=DET|PronType=Art",
"Mood=Ind|POS=VERB|Tense=Pres|VerbForm=Fin|Voice=Pass",
"Abbr=Yes|Case=Gen|POS=PROPN",
"Animacy=Hum|Case=Nom|Gender=Masc|Number=Sing|POS=PRON|Person=3|PronType=Prs",
"Degree=Cmp|POS=ADJ",
"POS=ADJ|VerbForm=Part",
"Gender=Neut|Number=Sing|POS=PRON|Poss=Yes|PronType=Prs",
"Abbr=Yes|POS=ADP",
"Definite=Ind|Gender=Neut|Number=Sing|POS=DET|PronType=Prs",
"Case=Gen|Definite=Def|Gender=Neut|Number=Plur|POS=NOUN",
"POS=AUX|VerbForm=Part",
"POS=PRON|PronType=Int",
"Gender=Fem|Number=Sing|POS=PRON|Poss=Yes|PronType=Prs",
"Number=Plur|POS=PRON|Person=3|PronType=Ind,Prs",
"Number=Plur|POS=DET|PronType=Ind",
"Degree=Pos|POS=ADJ",
"Animacy=Hum|Case=Nom|Number=Plur|POS=PRON|Person=1|PronType=Prs",
"POS=VERB|VerbForm=Inf|Voice=Pass",
"Definite=Ind|Gender=Fem|Number=Sing|POS=DET|PronType=Dem",
"Gender=Neut|Number=Sing|POS=DET|PronType=Ind",
"Animacy=Hum|Case=Acc|Gender=Masc|Number=Sing|POS=PRON|Person=3|PronType=Prs",
"Animacy=Hum|Case=Nom|Number=Sing|POS=PRON|Person=1|PronType=Prs",
"Number=Plur|POS=DET|Polarity=Neg|PronType=Neg",
"NumType=Card|POS=NUM",
"Gender=Masc|Number=Sing|POS=DET|PronType=Ind",
"POS=DET|PronType=Prs",
"Gender=Fem|Number=Sing|POS=DET|PronType=Ind",
"Case=Gen|Gender=Neut|POS=PROPN",
"Gender=Masc|Number=Sing|POS=DET|Polarity=Neg|PronType=Neg",
"Definite=Def|Number=Sing|POS=ADJ|VerbForm=Part",
"Gender=Fem,Masc|Number=Sing|POS=PRON|Person=3|PronType=Prs",
"POS=AUX|VerbForm=Inf",
"Case=Acc|Number=Plur|POS=PRON|Person=3|PronType=Prs",
"Case=Gen|Degree=Pos|Number=Plur|POS=ADJ",
"Number=Plur|POS=DET|PronType=Tot",
"Case=Gen|Gender=Masc|Number=Sing|POS=DET|PronType=Dem",
"Number=Plur|POS=DET|PronType=Prs",
"POS=SYM",
"Gender=Neut|NumType=Card|Number=Sing|POS=NUM",
"Animacy=Hum|Case=Nom|Number=Sing|POS=PRON|PronType=Prs",
"Definite=Ind|Gender=Masc|Number=Sing|POS=DET|PronType=Prs",
"Case=Gen|Definite=Ind|Gender=Masc|Number=Sing|POS=NOUN",
"Abbr=Yes|POS=ADV",
"Definite=Ind|Gender=Neut|Number=Sing|POS=DET|PronType=Dem",
"Gender=Masc|Number=Sing|POS=DET|PronType=Tot",
"Definite=Def|POS=DET|PronType=Prs",
"Animacy=Hum|Case=Nom|Gender=Fem|Number=Sing|POS=PRON|Person=3|PronType=Prs",
"Gender=Neut|POS=NOUN",
"Gender=Neut|Number=Sing|POS=DET|PronType=Int",
"Definite=Def|NumType=Card|POS=NUM",
"Mood=Imp|POS=VERB|VerbForm=Fin",
"Definite=Ind|Number=Plur|POS=NOUN",
"Gender=Neut|Number=Sing|POS=DET|PronType=Tot",
"Gender=Fem|Number=Sing|POS=DET|PronType=Tot",
"Animacy=Hum|Case=Acc|Number=Plur|POS=PRON|Person=1|PronType=Prs",
"Gender=Fem,Masc|Number=Sing|POS=PRON|Person=3|Polarity=Neg|PronType=Neg,Prs",
"Number=Plur|POS=PRON|Person=3|Polarity=Neg|PronType=Neg,Prs",
"Definite=Def|NumType=Card|Number=Sing|POS=NUM",
"Gender=Masc|NumType=Card|Number=Sing|POS=NUM",
"Definite=Ind|Gender=Masc|Number=Sing|POS=DET|PronType=Dem",
"Case=Gen|Definite=Def|Gender=Fem|Number=Plur|POS=NOUN",
"Case=Gen|Gender=Neut|Number=Sing|POS=DET|PronType=Dem",
"POS=SPACE",
"Animacy=Hum|Number=Sing|POS=PRON|PronType=Art,Prs",
"Mood=Imp|POS=AUX|VerbForm=Fin",
"Number=Plur|POS=PRON|Person=3|PronType=Prs,Tot",
"Number=Plur|POS=ADJ",
"Gender=Masc|POS=NOUN",
"Abbr=Yes|POS=NOUN",
"Case=Gen|Definite=Ind|Gender=Masc|Number=Plur|POS=NOUN",
"Gender=Neut|Number=Sing|POS=PRON|Person=3|PronType=Ind,Prs",
"POS=INTJ",
"Animacy=Hum|Case=Nom|Number=Sing|POS=PRON|Person=2|PronType=Prs",
"Animacy=Hum|Case=Acc|Number=Sing|POS=PRON|Person=1|PronType=Prs",
"Case=Gen|Definite=Def|Gender=Masc|Number=Plur|POS=NOUN",
"POS=ADJ",
"Animacy=Hum|Case=Acc|Gender=Fem|Number=Sing|POS=PRON|Person=3|PronType=Prs",
"Animacy=Hum|Case=Acc|Number=Sing|POS=PRON|Person=2|PronType=Prs",
"Definite=Def|Gender=Fem|Number=Sing|POS=NOUN",
"Number=Sing|POS=PRON|Polarity=Neg|PronType=Neg",
"Case=Gen|POS=NOUN",
"Definite=Ind|Number=Sing|POS=ADJ",
"Case=Gen|Gender=Masc|POS=PROPN",
"Animacy=Hum|Number=Plur|POS=PRON|PronType=Rcp",
"Case=Gen|Definite=Ind|Gender=Fem|Number=Sing|POS=NOUN",
"Number=Plur|POS=PRON|Person=3|PronType=Prs",
"Gender=Fem,Masc|Number=Sing|POS=PRON|Person=3|PronType=Ind,Prs",
"Definite=Ind|Gender=Fem|Number=Sing|POS=DET|PronType=Prs",
"Case=Gen|Definite=Def|Gender=Fem|Number=Sing|POS=NOUN",
"Gender=Fem|Number=Sing|POS=DET|PronType=Art",
"Case=Gen|Definite=Def|Degree=Pos|Number=Sing|POS=ADJ",
"Gender=Masc|Number=Sing|POS=DET|PronType=Int",
"NumType=Card|Number=Sing|POS=NUM",
"Animacy=Hum|Case=Acc|Number=Plur|POS=PRON|Person=2|PronType=Prs",
"Animacy=Hum|Case=Nom|Number=Plur|POS=PRON|Person=2|PronType=Prs",
"Case=Gen|Definite=Ind|Degree=Pos|Gender=Neut|Number=Sing|POS=ADJ",
"Degree=Sup|POS=ADJ",
"Animacy=Hum|POS=PRON|PronType=Int",
"POS=DET|PronType=Ind",
"Definite=Def|Number=Sing|POS=DET|PronType=Dem",
"Gender=Fem|POS=NOUN",
"Case=Gen|Number=Plur|POS=DET|PronType=Dem",
"Gender=Fem,Masc|Number=Sing|POS=PRON|Person=3|PronType=Prs,Tot",
"Case=Gen|Definite=Ind|Gender=Fem|Number=Plur|POS=NOUN",
"Gender=Neut|Number=Sing|POS=DET|Polarity=Neg|PronType=Neg",
"Number=Plur|POS=NOUN",
"POS=PRON|PronType=Prs",
"Case=Gen|Definite=Ind|Degree=Pos|Number=Sing|POS=ADJ",
"Definite=Ind|Number=Sing|POS=VERB|VerbForm=Part",
"Case=Gen|Definite=Def|Number=Sing|POS=ADJ|VerbForm=Part",
"Mood=Ind|POS=VERB|Tense=Past|VerbForm=Fin|Voice=Pass",
"Gender=Neut|Number=Sing|POS=DET|PronType=Dem,Ind",
"Animacy=Hum|POS=PRON|Poss=Yes|PronType=Int",
"Abbr=Yes|POS=ADJ",
"Case=Gen|Gender=Masc|Number=Sing|POS=DET|PronType=Art",
"Abbr=Yes|Definite=Def,Ind|Gender=Masc|Number=Sing|POS=NOUN",
"Case=Gen|Gender=Fem|Number=Sing|POS=DET|PronType=Dem",
"Number=Plur|POS=PRON|Poss=Yes|PronType=Rcp",
"Definite=Ind|Degree=Pos|POS=ADJ",
"Number=Plur|POS=DET|PronType=Art",
"Case=Gen|NumType=Card|Number=Plur|POS=NUM",
"Abbr=Yes|Definite=Def,Ind|Gender=Neut|Number=Plur,Sing|POS=NOUN",
"Case=Gen|Number=Plur|POS=DET|PronType=Tot",
"Abbr=Yes|Definite=Def,Ind|Gender=Masc|Number=Plur,Sing|POS=NOUN",
"Gender=Fem|Number=Sing|POS=DET|PronType=Int",
"Definite=Ind|Gender=Neut|Number=Sing|POS=ADJ",
"Case=Gen|Definite=Ind|Gender=Masc|Number=Sing|POS=DET|PronType=Dem",
"Gender=Fem|Number=Sing|POS=DET|PronType=Prs",
"Animacy=Hum|Case=Gen,Nom|Number=Sing|POS=PRON|PronType=Art,Prs",
"Definite=Def|Degree=Pos|Gender=Masc|Number=Sing|POS=ADJ",
"Animacy=Hum|Case=Gen|Number=Sing|POS=PRON|PronType=Art,Prs",
"Gender=Fem|NumType=Card|Number=Sing|POS=NUM",
"Definite=Ind|Gender=Masc|POS=NOUN",
"Definite=Def|Number=Plur|POS=NOUN",
"Number=Sing|POS=ADJ|VerbForm=Part",
"Definite=Ind|Gender=Masc|Number=Sing|POS=ADJ|VerbForm=Part",
"Abbr=Yes|Gender=Masc|POS=NOUN",
"Abbr=Yes|Case=Gen|POS=NOUN",
"Abbr=Yes|Mood=Ind|POS=VERB|Tense=Pres|VerbForm=Fin",
"Abbr=Yes|Degree=Pos|POS=ADJ",
"Case=Gen|Gender=Fem|POS=NOUN",
"Case=Gen|Degree=Cmp|POS=ADJ",
"Definite=Ind|Degree=Pos|Gender=Masc|Number=Sing|POS=ADJ",
"Gender=Masc|Number=Sing|POS=NOUN"
],
"parser":[
"ROOT",
"acl",
"acl:cleft",
"acl:relcl",
"advcl",
"advmod",
"amod",
"appos",
"aux",
"aux:pass",
"case",
"cc",
"ccomp",
"compound",
"compound:prt",
"conj",
"cop",
"csubj",
"dep",
"det",
"discourse",
"expl",
"flat:foreign",
"flat:name",
"iobj",
"mark",
"nmod",
"nsubj",
"nsubj:pass",
"nummod",
"obj",
"obl",
"orphan",
"parataxis",
"punct",
"xcomp"
],
"attribute_ruler":[
],
"ner":[
"DRV",
"EVT",
"GPE_LOC",
"GPE_ORG",
"LOC",
"MISC",
"ORG",
"PER",
"PROD"
]
},
"pipeline":[
"tok2vec",
"morphologizer",
"parser",
"lemmatizer",
"attribute_ruler",
"ner"
],
"components":[
"tok2vec",
"morphologizer",
"parser",
"lemmatizer",
"senter",
"attribute_ruler",
"ner"
],
"disabled":[
"senter"
],
"performance":{
"token_acc":0.9980718909,
"token_p":0.997080292,
"token_r":0.9953256895,
"token_f":0.9962022181,
"pos_acc":0.9728910684,
"morph_acc":0.9586576479,
"morph_micro_p":0.9759447339,
"morph_micro_r":0.9664952241,
"morph_micro_f":0.9711969942,
"morph_per_feat":{
"Definite":{
"p":0.9788408464,
"r":0.9654077351,
"f":0.972077885
},
"Gender":{
"p":0.9480072637,
"r":0.9408138101,
"f":0.944396839
},
"Number":{
"p":0.9767338471,
"r":0.9657850392,
"f":0.9712285872
},
"Mood":{
"p":0.9873641648,
"r":0.9769942486,
"f":0.9821518351
},
"Tense":{
"p":0.9880559085,
"r":0.9790984639,
"f":0.9835567923
},
"VerbForm":{
"p":0.9743422128,
"r":0.971642764,
"f":0.9729906161
},
"Degree":{
"p":0.9708130646,
"r":0.9598076262,
"f":0.9652789774
},
"PronType":{
"p":0.9911159263,
"r":0.9866264021,
"f":0.9888660685
},
"Poss":{
"p":0.9843137255,
"r":0.9920948617,
"f":0.9881889764
},
"Case":{
"p":0.9938478747,
"r":0.9839424142,
"f":0.9888703395
},
"Reflex":{
"p":1.0,
"r":1.0,
"f":1.0
},
"NumType":{
"p":1.0,
"r":0.9135802469,
"f":0.9548387097
},
"Polarity":{
"p":1.0,
"r":1.0,
"f":1.0
},
"Person":{
"p":0.9909824395,
"r":0.9909824395,
"f":0.9909824395
},
"Animacy":{
"p":0.9991467577,
"r":0.9974446337,
"f":0.9982949702
},
"Voice":{
"p":0.9776119403,
"r":0.9492753623,
"f":0.9632352941
},
"Abbr":{
"p":1.0,
"r":0.8375,
"f":0.9115646259
}
},
"sents_p":0.9511686144,
"sents_r":0.9460356995,
"sents_f":0.9485952133,
"dep_uas":0.8936788796,
"dep_las":0.8623767174,
"dep_las_per_type":{
"nmod":{
"p":0.7880715706,
"r":0.8132950349,
"f":0.8004846527
},
"nsubj":{
"p":0.8951937984,
"r":0.8896764253,
"f":0.8924265842
},
"flat:name":{
"p":0.8824476651,
"r":0.8824476651,
"f":0.8824476651
},
"root":{
"p":0.9084096586,
"r":0.9057700291,
"f":0.9070879235
},
"advmod":{
"p":0.8359636538,
"r":0.8387715931,
"f":0.8373652695
},
"mark":{
"p":0.9233333333,
"r":0.9243604004,
"f":0.9238465814
},
"advcl":{
"p":0.7,
"r":0.6843910806,
"f":0.6921075455
},
"obj":{
"p":0.8856799037,
"r":0.9131513648,
"f":0.8992058644
},
"case":{
"p":0.941588785,
"r":0.936,
"f":0.9387860748
},
"det":{
"p":0.9664769133,
"r":0.9670886076,
"f":0.9667826637
},
"obl":{
"p":0.7673975214,
"r":0.7285067873,
"f":0.747446611
},
"aux":{
"p":0.9425493716,
"r":0.9502262443,
"f":0.9463722397
},
"acl:relcl":{
"p":0.8013245033,
"r":0.8175675676,
"f":0.8093645485
},
"iobj":{
"p":0.7733333333,
"r":0.6590909091,
"f":0.7116564417
},
"xcomp":{
"p":0.7142857143,
"r":0.7142857143,
"f":0.7142857143
},
"cc":{
"p":0.9192546584,
"r":0.9128758674,
"f":0.9160541586
},
"conj":{
"p":0.7291518175,
"r":0.74291939,
"f":0.735971223
},
"amod":{
"p":0.9397664413,
"r":0.9261053907,
"f":0.932885906
},
"nummod":{
"p":0.9246031746,
"r":0.9510204082,
"f":0.9376257545
},
"compound:prt":{
"p":0.8104089219,
"r":0.7870036101,
"f":0.7985347985
},
"cop":{
"p":0.8710433763,
"r":0.8855780691,
"f":0.878250591
},
"appos":{
"p":0.3623188406,
"r":0.25,
"f":0.2958579882
},
"ccomp":{
"p":0.7992831541,
"r":0.8779527559,
"f":0.8367729831
},
"acl":{
"p":0.6929824561,
"r":0.6370967742,
"f":0.6638655462
},
"expl":{
"p":0.699751861,
"r":0.7621621622,
"f":0.7296248383
},
"csubj":{
"p":0.7542372881,
"r":0.6496350365,
"f":0.6980392157
},
"nsubj:pass":{
"p":0.8592964824,
"r":0.7916666667,
"f":0.8240963855
},
"aux:pass":{
"p":0.9075630252,
"r":0.9,
"f":0.9037656904
},
"acl:cleft":{
"p":0.3846153846,
"r":0.3703703704,
"f":0.3773584906
},
"orphan":{
"p":0.0,
"r":0.0,
"f":0.0
},
"dep":{
"p":0.0,
"r":0.0,
"f":0.0
},
"parataxis":{
"p":0.8219178082,
"r":0.6382978723,
"f":0.7185628743
},
"reparandum":{
"p":0.0,
"r":0.0,
"f":0.0
},
"compound":{
"p":0.6842105263,
"r":0.3170731707,
"f":0.4333333333
},
"flat:foreign":{
"p":0.72,
"r":0.5806451613,
"f":0.6428571429
},
"discourse":{
"p":0.6923076923,
"r":0.3913043478,
"f":0.5
},
"csubj:pass":{
"p":0.0,
"r":0.0,
"f":0.0
}
},
"lemma_acc":0.970965787,
"tag_acc":0.9728910684,
"ents_p":0.811827957,
"ents_r":0.8050947867,
"ents_f":0.8084473528,
"ents_per_type":{
"PER":{
"p":0.8961038961,
"r":0.9034369885,
"f":0.8997555012
},
"PROD":{
"p":0.624,
"r":0.4814814815,
"f":0.543554007
},
"GPE_LOC":{
"p":0.8127208481,
"r":0.888030888,
"f":0.8487084871
},
"ORG":{
"p":0.7962962963,
"r":0.8514851485,
"f":0.8229665072
},
"LOC":{
"p":0.7804878049,
"r":0.5871559633,
"f":0.6701570681
},
"DRV":{
"p":0.7571428571,
"r":0.6708860759,
"f":0.711409396
},
"EVT":{
"p":0.3846153846,
"r":0.5555555556,
"f":0.4545454545
},
"GPE_ORG":{
"p":0.6226415094,
"r":0.6,
"f":0.6111111111
}
},
"speed":11591.0424921653
},
"sources":[
{
"name":"UD Norwegian Bokmaal v2.8",
"url":"https://github.com/UniversalDependencies/UD_Norwegian-Bokmaal",
"license":"Public Domain (CC0)",
"author":"\u00d8vrelid, Lilja; J\u00f8rgensen, Fredrik; Hohle, Petter"
},
{
"name":"NorNE: Norwegian Named Entities (commit: bd311de5)",
"url":"https://github.com/ltgoslo/norne",
"license":"Public Domain (CC0)",
"author":"Language Technology Group (University of Oslo)"
},
{
"name":"Explosion fastText Vectors (cbow, OSCAR Common Crawl + Wikipedia)",
"url":"https://spacy.io",
"license":"CC0",
"author":"Explosion"
}
],
"requirements":[
]
}