Update app.py

app.py

@@ -41,7 +41,7 @@ pipe = DiffusionPipeline.from_pretrained(
 
 # 4) Memory savings hooks (all on your single GPU + CPU offload)
 pipe.enable_attention_slicing()                     # slice big attention maps
-pipe.enable_vae_slicing()                           # slice VAE decode
+pipe.vae.enable_slicing()                          # slice VAE decode
 pipe.enable_xformers_memory_efficient_attention()   # if xformers is installed
 pipe.enable_model_cpu_offload()                     # offload idle submodules to CPU