Update models.py

models.py

@@ -398,6 +398,11 @@ def setup_scheduler(pipe):
 
 def optimize_pipeline(pipe):
     """Apply optimizations to pipeline."""
+    
+    # Enable CPU offloading for VRAM-constrained environments
+    print("  [OK] Enabling model CPU offloading...")
+    pipe.enable_model_cpu_offload()
+    
     # Try to enable xformers
     if device == "cuda":
         try: