Update app.py

app.py

@@ -7,8 +7,8 @@ from peft import (
     prepare_model_for_kbit_training,
     get_peft_model,
 )
-model_name = "google/gemma-2-2b-it"
-lora_model_name="Anlam-Lab/gemma-2-2b-it-anlamlab-SA-Chatgpt4mini"
+model_name = "meta-llama/Llama-3.2-1B"
+lora_model_name="Anlam-Lab/Llama-3.2-1B-it-anlamlab-SA-Chatgpt4mini"
 
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", torch_dtype=torch.float16)
@@ -32,7 +32,7 @@ def generate_response(input_text):
         )
     
     response = tokenizer.decode(outputs[0])
-    return response.split("<start_of_turn>model\n")[1].split("<end_of_turn>")[0]
+    return response.split("<|end_header_id|>")[-1].split("<|eot_id|>")[0]
 
 iface = gr.Interface(
     fn=generate_response,