Update app.py

app.py

@@ -137,15 +137,15 @@ def load_models_and_documents():
             type_v=GGML_TYPE_I8
         )
 
-        dense_model = OptimumEncoder(
-            device="cuda",
-            cache_dir=os.getenv('HF_HOME')
-        )
+        #dense_model = OptimumEncoder(
+            #device="cuda",
+            #cache_dir=os.getenv('HF_HOME')
+        #)
         
         sparse_model = SparseTextEmbedding(
             'Qdrant/bm42-all-minilm-l6-v2-attentions',
             cache_dir=os.getenv('HF_HOME'),
-            providers=['CUDAExecutionProvider']
+            providers=['CPUExecutionProvider']
         )
 
         data_dir = os.path.join(os.getenv('HF_HOME'), 'nltk_data')