Update app.py

app.py

@@ -19,20 +19,22 @@ CORS(app)
 
 logger = get_logger()
 logger.warning('Entering application')
-os.environ["TOKENIZERS_PARALLELISM"] = "true"
+os.environ["TOKENIZERS_PARALLELISM"] = "false"
 
 def load_model():
     logger.warning('Entering load transformer')
-    interpreter = tf.lite.Interpreter(model_path=os.path.join("models/news_classification_hf_distilbert.tflite"))
-    with open("models/news_classification_labelencoder.bin", "rb") as model_file_obj:
+    with open("models/label_encoder.bin", "rb") as model_file_obj:
         label_encoder = cloudpickle.load(model_file_obj)
-        
-    model_checkpoint = "distilbert-base-uncased"
-    tokenizer = DistilBertTokenizerFast.from_pretrained(model_checkpoint)
+
+    with open("models/calibrated_model.bin", "rb") as model_file_obj:
+        calibrated_model = cloudpickle.load(model_file_obj)
+
+    tflite_model_path = os.path.join("models", "model.tflite")
+    calibrated_model.estimator.tflite_model_path = tflite_model_path
     logger.warning('Exiting load transformer')
-    return interpreter, label_encoder, tokenizer
+    return calibrated_model, label_encoder
 
-interpreter, label_encoder, tokenizer = load_model()
+calibrated_model, label_encoder = load_model()
 vectorizer = TextVectorizer()
 collection = get_milvus_collection()
 sent_model, ce_model = load_sentence_transformer()
@@ -50,8 +52,8 @@ def update_news():
         prediction_db_write = DBWrite(db_type="prediction")
         old_news = db_read.read_news_from_db()
         new_news = get_news()
-        news_df, prediction_df, is_db_updation_required = predict_news_category_similar_news(old_news, new_news, interpreter, label_encoder, 
-                                                                              tokenizer, collection, vectorizer, sent_model, ce_model)
+        news_df, prediction_df, is_db_updation_required = predict_news_category_similar_news(old_news, new_news, calibrated_model, label_encoder,
+                                                                                             collection, vectorizer, sent_model, ce_model)
         if news_df is None:
             raise Exception('Could not generate category predictions. Aborting the database insertion. No new articles are inserted into the collection.')