Update app.py

app.py

@@ -13,7 +13,12 @@ import pkg_resources
 import folium
 import country_converter as coco
 import time
+import gc
 
+def clear_memory():
+    if torch.cuda.is_available():
+        torch.cuda.empty_cache()
+    gc.collect()
     
 current_dir = os.path.dirname(os.path.abspath(__file__))
 font_path = os.path.join(current_dir, "ArabicR2013-J25x.ttf")
@@ -33,7 +38,7 @@ ARABIC_STOP_WORDS = {
     'منذ', 'نفس', 'حيث', 'هناك', 'جدا', 'ذات', 'ضمن', 'انه', 'لدى',
     'عليه', 'مثل', 'أما', 'لدي', 'فيه', 'كلم', 'لكن', 'ايضا', 'لازم',
      'يجب', 'صار', 'صارت', 'ضد', 'يا', 'لا', 'اما',
-    'بها', 'ان', 'به', 'الي', 'لما', 'انا', 'اليك', 'لي', 'لك', 'قلت',
+    'بها', 'ان', 'به', 'الي', 'لما', 'انا', 'اليك', 'لي', 'لك','اذا','بلا','او','لديك','لديه', 'قلت',
 
     # Middle group prefixed with "و"
     'وثم', 'وأو', 'ول', 'وب', 'وك', 'ولل', 'وال', 
@@ -402,6 +407,10 @@ def process_and_summarize(df, bert_tokenizer, bert_model, emotion_classifier, to
         all_emotions = []
         
         embeddings = []
+
+        clear_memory()
+
+        
         for i, text in enumerate(texts):
             try:
                 embedding = get_embedding_for_text(text, bert_tokenizer, bert_model)
@@ -413,6 +422,9 @@ def process_and_summarize(df, bert_tokenizer, bert_model, emotion_classifier, to
             except Exception as e:
                 st.warning(f"Error generating embedding for text {i+1} in {country}: {str(e)}")
                 continue
+            if i % 10 == 0:
+                clear_memory()
+            
             progress = (i + 1) / len(texts) * 0.4
             progress_bar.progress(progress, text=f"Generated embeddings for {i+1}/{len(texts)} poems...")
         
@@ -420,9 +432,13 @@ def process_and_summarize(df, bert_tokenizer, bert_model, emotion_classifier, to
             texts = texts[:len(embeddings)]
         embeddings = np.array(embeddings)
         
+        clear_memory()
+
         for i, text in enumerate(texts):
             emotion = classify_emotion(text, emotion_classifier)
             all_emotions.append(emotion)
+            if i % 10 == 0:
+                clear_memory()
             progress = 0.4 + ((i + 1) / len(texts) * 0.3)
             progress_bar.progress(progress, text=f"Classified emotions for {i+1}/{len(texts)} poems...")