Update app.py

app.py

@@ -15,7 +15,9 @@ vae = AutoencoderKLWan.from_pretrained(model_id, subfolder="vae", torch_dtype=to
 pipe = WanImageToVideoPipeline.from_pretrained(
     model_id, vae=vae, image_encoder=image_encoder, torch_dtype=torch.float16
 )
-pipe.to("cuda" if torch.cuda.is_available() else "cpu")
+device = "cuda" if torch.cuda.is_available() else "cpu"
+pipe.to(device)
+pipe.enable_model_cpu_offload()
 
 # --- Helper Functions ---
 def aspect_ratio_resize(image, pipe, max_area=720 * 1280):
@@ -50,8 +52,7 @@ def infer(first_image, last_image, prompt, guidance=5.5, frames=25):
 
     # Run pipeline
     output = pipe(
-        image=first_image,
-        last_image=last_image,
+        image=[first_image, last_image],
         prompt=prompt,
         height=height,
         width=width,
@@ -80,4 +81,4 @@ demo = gr.Interface(
 )
 
 if __name__ == "__main__":
-    demo.launch(show_api=True)
+    demo.launch(show_api=True, delete_cache=(60, 60))