Update app.py

app.py

@@ -79,7 +79,7 @@ def generate_response(system_instruction, user_input):
     with torch.no_grad():
         generated_ids = model.generate(model_inputs, max_new_tokens=1000, do_sample=True)
     # Extract only the bot's response, omitting the prompt
-    response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0].split("\n")[-1].strip()
+    response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0].split("\n")[-2:]
 
     return response