Update app.py

app.py

@@ -10,7 +10,7 @@ model_name = "thedeba/debai"  # HF Hub model path
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 model = AutoModelForCausalLM.from_pretrained(model_name)
 device = "cpu"  # Spaces free tier uses CPU; you can switch to "cuda" if GPU granted
-model.to(device)
+#model.to(device)
 
 # -------------------------------
 # FastAPI setup