hiroshi-matsuda-rit's picture
Update spaCy pipeline
91bd98b
{
"lang":"ja",
"name":"gsd_bert_wwm_unidic_lite",
"version":"3.1.1",
"description":"Japanese transformer pipeline (bert-base). Components: transformer, parser, ner.",
"author":"Megagon Labs Tokyo.",
"email":"ginza@megagon.ai",
"url":"https://github.com/megagonlabs/UD_japanese_GSD",
"license":"CC BY-SA 4.0",
"spacy_version":">=3.1.0,<3.2.0",
"spacy_git_version":"530b5d72f",
"vectors":{
"width":0,
"vectors":0,
"keys":0,
"name":null
},
"labels":{
"transformer":[
],
"parser":[
"ROOT",
"acl",
"advcl",
"advmod",
"amod",
"aux",
"case",
"cc",
"ccomp",
"compound",
"cop",
"csubj",
"dep",
"det",
"dislocated",
"fixed",
"mark",
"nmod",
"nsubj",
"nummod",
"obj",
"obl",
"punct"
],
"ner":[
"CARDINAL",
"DATE",
"EVENT",
"FAC",
"GPE",
"LANGUAGE",
"LAW",
"LOC",
"MONEY",
"MOVEMENT",
"NORP",
"ORDINAL",
"ORG",
"PERCENT",
"PERSON",
"PET_NAME",
"PHONE",
"PRODUCT",
"QUANTITY",
"TIME",
"TITLE_AFFIX",
"WORK_OF_ART"
]
},
"pipeline":[
"transformer",
"parser",
"ner"
],
"components":[
"transformer",
"parser",
"ner"
],
"disabled":[
],
"performance":{
"dep_uas":0.9367795389,
"dep_las":0.926075995,
"dep_las_per_type":{
"cc":{
"p":0.8863636364,
"r":0.8125,
"f":0.847826087
},
"compound":{
"p":0.9503214494,
"r":0.916572717,
"f":0.9331420373
},
"obl":{
"p":0.8710493047,
"r":0.8601747815,
"f":0.8655778894
},
"case":{
"p":0.9870030581,
"r":0.9810030395,
"f":0.9839939024
},
"dislocated":{
"p":0.6,
"r":0.6923076923,
"f":0.6428571429
},
"nsubj":{
"p":0.8875739645,
"r":0.8637236084,
"f":0.8754863813
},
"nmod":{
"p":0.9405063291,
"r":0.869005848,
"f":0.903343465
},
"root":{
"p":0.9119850187,
"r":0.9605522682,
"f":0.9356388088
},
"aux":{
"p":0.9766573296,
"r":0.9712163417,
"f":0.9739292365
},
"advcl":{
"p":0.7825112108,
"r":0.7842696629,
"f":0.7833894501
},
"mark":{
"p":0.979757085,
"r":0.968,
"f":0.9738430584
},
"fixed":{
"p":0.9694793537,
"r":0.9818181818,
"f":0.9756097561
},
"acl":{
"p":0.8984198646,
"r":0.8747252747,
"f":0.8864142539
},
"obj":{
"p":0.9541284404,
"r":0.9425981873,
"f":0.9483282675
},
"nummod":{
"p":0.9934640523,
"r":0.899408284,
"f":0.9440993789
},
"advmod":{
"p":0.7647058824,
"r":0.7428571429,
"f":0.7536231884
},
"amod":{
"p":0.935483871,
"r":0.7837837838,
"f":0.8529411765
},
"cop":{
"p":0.9759036145,
"r":0.9418604651,
"f":0.9585798817
},
"ccomp":{
"p":0.9523809524,
"r":0.9090909091,
"f":0.9302325581
},
"det":{
"p":1.0,
"r":0.9811320755,
"f":0.9904761905
},
"csubj":{
"p":0.7142857143,
"r":0.8333333333,
"f":0.7692307692
},
"dep":{
"p":0.4,
"r":0.2857142857,
"f":0.3333333333
}
},
"sents_p":0.9201520913,
"sents_r":0.9546351085,
"sents_f":0.9370764763,
"ents_f":0.840432295,
"ents_p":0.8496143959,
"ents_r":0.8314465409,
"ents_per_type":{
"DATE":{
"p":0.9814814815,
"r":0.9724770642,
"f":0.9769585253
},
"ORG":{
"p":0.78125,
"r":0.7299270073,
"f":0.7547169811
},
"TITLE_AFFIX":{
"p":0.8518518519,
"r":0.7666666667,
"f":0.8070175439
},
"PERSON":{
"p":0.9333333333,
"r":0.9064748201,
"f":0.9197080292
},
"GPE":{
"p":0.780952381,
"r":0.8723404255,
"f":0.824120603
},
"PRODUCT":{
"p":0.6285714286,
"r":0.5238095238,
"f":0.5714285714
},
"TIME":{
"p":0.6666666667,
"r":1.0,
"f":0.8
},
"QUANTITY":{
"p":0.8648648649,
"r":0.9696969697,
"f":0.9142857143
},
"NORP":{
"p":0.8846153846,
"r":0.71875,
"f":0.7931034483
},
"ORDINAL":{
"p":0.7391304348,
"r":0.7727272727,
"f":0.7555555556
},
"WORK_OF_ART":{
"p":0.6666666667,
"r":0.7058823529,
"f":0.6857142857
},
"PERCENT":{
"p":1.0,
"r":0.5714285714,
"f":0.7272727273
},
"CARDINAL":{
"p":0.0,
"r":0.0,
"f":0.0
},
"EVENT":{
"p":0.9565217391,
"r":0.8461538462,
"f":0.8979591837
},
"FAC":{
"p":0.8055555556,
"r":0.7837837838,
"f":0.7945205479
},
"LOC":{
"p":0.8888888889,
"r":0.8,
"f":0.8421052632
},
"MOVEMENT":{
"p":0.6,
"r":0.6,
"f":0.6
},
"LAW":{
"p":1.0,
"r":1.0,
"f":1.0
},
"MONEY":{
"p":1.0,
"r":1.0,
"f":1.0
},
"LANGUAGE":{
"p":1.0,
"r":1.0,
"f":1.0
}
},
"tag_acc":0.0,
"transformer_loss":288.6167381342,
"parser_loss":13062.4862750822,
"ner_loss":139.9335659693
},
"sources":[
{
"name":"UD_Japanese-GSD",
"url":"https://github.com/UniversalDependencies/UD_Japanese-GSD",
"license":"CC BY-SA 4.0"
},
{
"name":"UD_Japanese-GSD r2.8+NE",
"url":"https://github.com/megagonlabs/UD_Japanese-GSD/releases/tag/r2.8-NE",
"license":"CC BY-SA 4.0"
},
{
"name":"SudachiDict_core",
"url":"https://github.com/WorksApplications/SudachiDict",
"license":"Apache License 2.0"
},
{
"name":"cl-tohoku/bert-base-japanese-whole-word-masking",
"url":"https://huggingface.co/cl-tohoku/bert-base-japanese-whole-word-masking",
"license":"CC BY-SA 3.0"
},
{
"name":"unidic_lite",
"url":"https://github.com/polm/unidic-lite",
"license":"individually defined open software license"
}
],
"parent_package":"spacy",
"requirements":[
"sudachipy>=0.5.2",
"sudachidict_core>=20210608",
"spacy-transformers>=1.0.2,<1.1.0",
"fugashi",
"unidic_lite"
]
}