generate_response fix

app.py

@@ -34,8 +34,10 @@ def get_model_and_tokenizer(model_id):
 def generate_response(user_input, model_id):
     prompt = formatted_prompt(user_input)
     
-    # Load the model and tokenizer if they are not already loaded
-    if model is None or tokenizer is None:
+    global model, tokenizer
+
+    # Load the model and tokenizer if they are not already loaded or if the model_id has changed
+    if model is None or tokenizer is None or (model.config._name_or_path != model_id):
         get_model_and_tokenizer(model_id)  # Load model and tokenizer
 
     # Prepare the input tensors
@@ -44,7 +46,6 @@ def generate_response(user_input, model_id):
     generation_config = GenerationConfig(
         max_new_tokens=100,
         min_length=5,
-        #temperature=0.7, 
         do_sample=False,
         num_beams=1,
         pad_token_id=tokenizer.eos_token_id,
@@ -59,7 +60,7 @@ def generate_response(user_input, model_id):
     except Exception as e:
         print(f"Error generating response: {e}")
         return "Error generating response."
-
+    
 def formatted_prompt(question) -> str:
     return f"<|im_start|>user\n{question}<|im_end|>\n<|im_start|>assistant:"