Fix CUDA generator error by moving pipeline to GPU

Add pipe.to("cuda") inside GPU-decorated function to ensure
pipeline is on CUDA before using CUDA generator

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

app.py

@@ -54,7 +54,10 @@ def generate_interpolation(
             f"Parameters: frames={num_frames}, steps={num_inference_steps}, guidance={guidance_scale}"
         )
 
-        # Generate video - ZeroGPU handles device placement automatically
+        # Move pipeline to CUDA within the GPU-decorated function
+        pipe.to("cuda")
+
+        # Generate video
         generator = torch.Generator(device="cuda").manual_seed(seed)
 
         video = pipe(