Merge branch 'mikelandis01:master' into master

Author: lukeblevins

models/named-entity-recognition/gen_ner_data.py

@@ -198,7 +198,7 @@ def main():
     root_dir = Path(args.root_dir)
 
     spacy.prefer_gpu()
-    nlp = spacy.load("en_core_web_trf")
+    nlp = spacy.load("en_core_web_md")
     nlp.add_pipe("align_noun_phrases", name="align_noun_phrases", after="ner")
 
     doc_bin = spacy.tokens.DocBin(attrs=["ENT_IOB", "ENT_TYPE"])

poligrapher/utils.py

@@ -204,7 +204,7 @@ def setup_nlp_pipeline(ner_path: str):
     else:
         our_ner = spacy.load(ner_path)
 
-    nlp = spacy.load("en_core_web_trf", disable=["ner"])
+    nlp = spacy.load("en_core_web_md", disable=["ner"])
 
     # Disable spaCy's NER and use our NER
     our_ner.replace_listeners("transformer", "ner", ["model.tok2vec"])

requirements.txt

@@ -30,8 +30,8 @@ spacy-alignments==0.9.1
 spacy-transformers==1.3.5
 tldextract==5.0.1
 tokenizers==0.15.2
-torch==2.5.0
-transformers==4.36.2
+torch==2.2.2
+transformers>=4.36.2
 unidecode==1.3.8
 url-normalize==1.4.3
 werkzeug==2.3.7