[FIX] add doc2vec_workers

experimentsClassicClassifiers.py

@@ -91,7 +91,7 @@ for columnInput in [columnText]:
     features_techniques = [
     ('counter',  extractor.count_vect(max_df = vectorization_max_df, min_df = vectorization_min_df, numberOfFeatures = vectorization_numberOfFeatures )),
     ('tf_idf',  extractor.tf_idf(max_df = vectorization_max_df, min_df = vectorization_min_df, numberOfFeatures = vectorization_numberOfFeatures)),
-    ('doc2vec',  extractor.doc2vec(max_epochs, doc2vec_vec_size, doc2vec_min_count ,  doc2vec_dm))]
+    ('doc2vec',  extractor.doc2vec(max_epochs, doc2vec_vec_size, doc2vec_min_count ,  doc2vec_dm, doc2vec_workers))]
 
 
 

features_extractor.py

@@ -61,7 +61,7 @@ class feature_extractor:
 
 
 
-    def doc2vec(self, max_epochs, doc2vec_vec_size, doc2vec_min_count ,  doc2vec_dm):
+    def doc2vec(self, max_epochs, doc2vec_vec_size, doc2vec_min_count ,  doc2vec_dm, doc2vec_workers):
 
         nlp = spacy.load("fr_core_news_sm")
         stopWords = set(stopwords.words('french'))
@@ -84,7 +84,7 @@ class feature_extractor:
         #Tag test set
         tagged_test = [TaggedDocument(words=tokenize_fr_text(_d), tags = [str(i)]) for i, _d in enumerate(self.docs_test)]
 
-        model = Doc2Vec(vector_size=doc2vec_vec_size, min_count = doc2vec_min_count, dm = doc2vec_dm)
+        model = Doc2Vec(vector_size=doc2vec_vec_size, min_count = doc2vec_min_count, dm = doc2vec_dm, workers = doc2vec_workers)
         model.build_vocab(tagged_tr)
         model.train(tagged_tr, total_examples=model.corpus_count, epochs = max_epochs)
 

settings.conf

@@ -6,5 +6,5 @@ doc2vec_vec_size = 700
 max_epochs = 10
 doc2vec_min_count = 12
 doc2vec_dm = 0
-doc2vec_workers = 4
+doc2vec_workers = 8
 min_word_per_article = 25