Update app.py

app.py

@@ -33,11 +33,6 @@ device = "CPU"
 default_model_name = "Qwen3-0.6B-int4-ov"  # Choose a default model
 model_path = model_name_to_id[default_model_name].split("/")[-1]
 
-pipe = ov_genai.LLMPipeline(model_path, device)
-tokenizer = pipe.get_tokenizer()
-tokenizer.set_chat_template(tokenizer.chat_template)
-
-
 def generate_response(prompt, model_name):
     global pipe, tokenizer  # Access the global variables
 
@@ -45,11 +40,11 @@ def generate_response(prompt, model_name):
     model_id = model_name_to_id[model_name]
     new_model_path = model_id.split("/")[-1]
 
-    if pipe.model_name != new_model_path: # Assuming the LLMPipeline has a model_name property
-        print(f"Switching to model: {model_name}")
-        pipe = ov_genai.LLMPipeline(new_model_path, device)
-        tokenizer = pipe.get_tokenizer()
-        tokenizer.set_chat_template(tokenizer.chat_template)
+    
+    print(f"Switching to model: {model_name}")
+    pipe = ov_genai.LLMPipeline(new_model_path, device)
+    tokenizer = pipe.get_tokenizer()
+    tokenizer.set_chat_template(tokenizer.chat_template)
 
 
     try: