Update app.py

app.py

@@ -10,6 +10,11 @@ WATERMARK_TEXT = "SelamGPT"
 DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
 TORCH_DTYPE = torch.float16 if DEVICE == "cuda" else torch.float32
 
+try:
+    pipe.enable_xformers_memory_efficient_attention()
+except:
+    print("Xformers not available, using default attention")
+
 # ===== MODEL LOADING WITH GRADIO CACHE =====
 @gr.Cache()  # Now works in Gradio 5.x
 def load_model():