Update app.py

app.py

@@ -39,6 +39,13 @@ pipe.load_lora_weights(
     weight_name="40kHalf.safetensors"
 )
 
+# Enable CPU offload
+pipe.enable_model_cpu_offload()
+
+# Enable memory efficient attention if available
+if hasattr(pipe, "enable_xformers_memory_efficient_attention"):
+    pipe.enable_xformers_memory_efficient_attention()
+
 def get_canny_image(image, low_threshold=100, high_threshold=200):
     if isinstance(image, Image.Image):
         image = np.array(image)