Error on spacy.load(): "writing directly to language.factories isn't needed anymore in spacy v3. instead..." #7792
-
I upgraded to spacy v3 this morning. I'm on 3.0.5. I also downloaded the latest version of 'en_core_web_lg' I used the config file quick start for creating a config file but I get this error with writing directly to I haven't modified the base_config.cfg file from the quick start at all. This is an auto-generated partial config. To use it with 'spacy train' [system] [nlp] [components] [components.tok2vec] [components.tok2vec.model] [components.tok2vec.model.embed] [components.tok2vec.model.encode] [components.ner] [components.ner.model] [components.ner.model.tok2vec] [corpora] [corpora.train] [corpora.dev] [training] [training.optimizer] [training.batcher] [training.batcher.size] [initialize] |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
Update: this issue was caused by the spacy-dbpedia-spotlight library when I run spacy.load(model_name). I uninstalled this package and it worked ~\AppData\Local\Continuum\anaconda3\lib\site-packages\spacy_dbpedia_spotlight_init_.py in |
Beta Was this translation helpful? Give feedback.
Update: this issue was caused by the spacy-dbpedia-spotlight library when I run spacy.load(model_name). I uninstalled this package and it worked
~\AppData\Local\Continuum\anaconda3\lib\site-packages\spacy_dbpedia_spotlight_init_.py in
5 version = util.pkg_meta["version"]
6
----> 7 Language.factories['overwrite_vectors'] = lambda nlp, **cfg: EntityLinker(nlp, **cfg)
8
9 def load(language_code, nlp=None):