Update app.py

app.py

@@ -268,6 +268,11 @@ def initialize_fn():
     if audio_decoder is not None:
         return
     
+    print("Initializing models and managers...")
+    
+    # Initialize session manager first
+    session_manager = SessionManager()
+    
     model_worker = ModelWorker(args.model_path, device)
     glm_tokenizer = model_worker.glm_tokenizer
     
@@ -282,10 +287,11 @@ def initialize_fn():
     feature_extractor = WhisperFeatureExtractor.from_pretrained(args.tokenizer_path)
     
     embedding_model = create_embedding_model(Embedding_Model)
-    session_manager = SessionManager()
     vector_store_manager = VectorStoreManager(session_manager, embedding_model)
     
     whisper_transcribe_model = whisper.load_model("base")
+    
+    print("Initialization complete.")
 
 def clear_fn():
     return [], [], '', '', '', None, None
@@ -455,10 +461,13 @@ if __name__ == "__main__":
     # Configuration
     Embedding_Model = 'intfloat/multilingual-e5-large-instruct'
 
+    # Initialize models first
+    initialize_fn()
+
     # Create Gradio interface
     with gr.Blocks(title="GLM-4-Voice Demo", fill_height=True) as demo:
-        # Add session state
-        session_id = gr.State(lambda: session_manager.create_session())
+        # Now session_manager is initialized
+        session_id = gr.State(session_manager.create_session())
         
         with gr.Row():
             # Left column for chat interface