Hanlp的load参数疑问

github给的例子 HanLP = hanlp.load(hanlp.pretrained.mtl.CLOSE_TOK_POS_NER_SRL_DEP_SDP_CON_ELECTRA_SMALL_ZH) ,经查看发现这里面的参数有多种,有连接介绍每个参数的意思么?或者这几个参数之间的区别点在哪里?
OPEN_TOK_POS_NER_SRL_DEP_SDP_CON_ELECTRA_SMALL_ZH
OPEN_TOK_POS_NER_SRL_DEP_SDP_CON_ELECTRA_BASE_ZH
CLOSE_TOK_POS_NER_SRL_DEP_SDP_CON_ELECTRA_SMALL_ZH
CLOSE_TOK_POS_NER_SRL_DEP_SDP_CON_ELECTRA_BASE_ZH
UD_ONTONOTES_TOK_POS_LEM_FEA_NER_SRL_DEP_SDP_CON_MT5_SMALL
UD_ONTONOTES_TOK_POS_LEM_FEA_NER_SRL_DEP_SDP_CON_XLMR_BASE
NPCMJ_UD_KYOTO_TOK_POS_CON_BERT_BASE_CHAR_JA

您好,很高兴回答您的问题。

  • 参数释义如下:

TOK :tokenization(词语切分)
POS :Part-of-speech tagging(词性标注)
NER :Named Entity Recognition( 命名实体识别)
SRL :Semantic Role Labeling (语义角色标注)
DEP :Dependency Parsing(依存句法分析)
SDP :Semantic Dependency Parsing(语义依存分析)
CON:Constituency Parsing(短语成分分析)
LEM : Lemmatization(词干提取)
FEA :Features of Universal Dependencies(词法语法特征)
AMR: Abstract Meaning Representation(抽象意义表示)

如有对此后续问题也欢迎在此贴提问。 :rose: