Update app.py

app.py

@@ -114,6 +114,9 @@ def load_and_prepare_model():
     pipe.to(device=device, dtype=torch.bfloat16)
     pipe.vae = vaeXL.to(device) #.to('cpu') #.to(torch.bfloat16)
     
+    # Enable attention slicing
+    pipe.enable_attention_slicing()
+
     pipe.unet.set_attn_processor(AttnProcessor2_0())
     pipe.vae.set_default_attn_processor()
     return pipe