Update app.py

app.py

@@ -11,15 +11,17 @@ base_model = "meta-llama/Meta-Llama-3-8B-Instruct"
 
 llama_model = transformers.AutoModelForCausalLM.from_pretrained(base_model)
 
-peft_model = PeftModel.from_pretrained(llama_model, model_id)
 
 pipeline = transformers.pipeline(
     "text-generation",
-    model=peft_model,
+    model=llama_model,
+    tokenizer=base_model,
     model_kwargs={"torch_dtype": torch.bfloat16},
     device="cuda",
 )
 
+pipeline.model = PeftModel.from_pretrained(llama_model, model_id)
+
 def chat_function(message, history, system_prompt, max_new_tokens, temperature):
     messages = [{"role":"system","content":system_prompt},
                 {"role":"user", "content":message}]