Update app.py

app.py

@@ -2,8 +2,14 @@ import gradio as gr
 from llama_cpp import Llama
 
 # 1. Path to your GGUF file inside the Space repository
-MODEL_PATH = "simonper/fine-tuned-gguf-modal1/Llama-3.2-1B.Q8_0.gguf"   # <- change if your file is named differently
+#MODEL_PATH = "simonper/fine-tuned-gguf-modal1/Llama-3.2-1B.Q8_0.gguf"   # <- change if your file is named differently
 
+llm = Llama.from_pretrained(
+	repo_id="simonper/fine-tuned-gguf-modal1",
+	filename="Llama-3.2-1B.Q8_0.gguf",
+)
+
+"""
 # 2. Load the GGUF model once at startup
 llm = Llama(
     model_path=MODEL_PATH,
@@ -11,7 +17,7 @@ llm = Llama(
     n_threads=8,       # tweak based on CPU in the Space
     n_gpu_layers=0,    # 0 = pure CPU, >0 if GPU layers are available
 )
-
+"""
 
 def build_prompt(system_message: str, history: list[dict], user_message: str) -> str:
     """