minor updates

dev_set_training.csv

@@ -8,7 +8,7 @@ ADDRESS1|ADDRESS2|ARE_SAME
 17752 MAIN ST, HANOVER, PA 23393|177 52 MAIN ST, HANOVER, PA 23393|1
 17752 MAIN ST, HANOVER, PA 23393|177 52 MAIN STREET, HANOVER, PA 23393|1
 217-12 LOUDON RD, CONCORD, NH 03301|217-22 LOUDON RD, CONCORD, NH 03301|0
-2575 US HWY 43, ST 3-A, WINFIELD, AL 35594|2575 US HWY 43, ST 3A, WINFIELD, AL 35594|1
+2575 US HWY 43, ST 3-A, WINFIELD, AL 35594|25-75 US HWY 43, STREET 3A, WINFIELD, AL 35594|1
 2575 US HWY 43, ST 3-A, WINFIELD, AL 35594|2575 US HWY 43, ST 3B, WINFIELD, AL 35594|0
 440 TECHNOLOGY CENTER DRIVE, BOSTON, MA 10034|200 TECHNOLOGY CENTER DRIVE, BOSTON, MA 10034|0
 440 TECHNOLOGY CENTER DRIVE, BOSTON, MA 10034|440 TECHNOLOGY CENTER DR., BOSTON, MA 10034|1

eval.py

@@ -0,0 +1,35 @@
+from sentence_transformers import LoggingHandler, SentenceTransformer, evaluation
+from sentence_transformers.readers import InputExample
+import csv
+import logging
+
+#### Just some code to print debug information to stdout
+logging.basicConfig(format='%(asctime)s - %(message)s',
+                    datefmt='%Y-%m-%d %H:%M:%S',
+                    level=logging.INFO,
+                    handlers=[LoggingHandler()])
+logger = logging.getLogger(__name__)
+#### /print debug information to stdout
+
+model_name = 'sentence-transformers/paraphrase-albert-base-v2'
+
+
+model_sbert = SentenceTransformer(model_name)
+
+dev_sentences1 = []
+dev_sentences2 = []
+dev_labels = []
+with open( "dev_set_training.csv", encoding='utf8') as fIn:
+    reader = csv.DictReader(fIn, delimiter='|', quoting=csv.QUOTE_NONE)
+    for row in reader:
+        dev_sentences1.append(row['ADDRESS1'])
+        dev_sentences2.append(row['ADDRESS2'])
+        dev_labels.append(int(row['ARE_SAME']))
+
+
+
+binary_acc_evaluator = evaluation.BinaryClassificationEvaluator(dev_sentences1, dev_sentences2, dev_labels)
+
+binary_acc_evaluator(model_sbert)
+
+print(binary_acc_evaluator)

train.py

@@ -26,7 +26,8 @@ logger = logging.getLogger(__name__)
 #As base model, we use DistilBERT-base that was pre-trained on NLI and STSb data
 model = SentenceTransformer('sentence-transformers/paraphrase-albert-base-v2')
 num_epochs = 12
-train_batch_size = 14
+# Smaller is generally better more accurate results.
+train_batch_size = 5
 
 #As distance metric, we use cosine distance (cosine_distance = 1-cosine_similarity)
 distance_metric = losses.SiameseDistanceMetric.COSINE_DISTANCE