Update app.py

app.py

@@ -56,7 +56,7 @@ random_seed = 42
 video_length = 201
 W = 1024
 H = W
-device = "cpu" if torch.cuda.is_available() else "cpu"
+device = "cpu"
 
 def get_pipe_image_and_video_predictor():
     vae = AutoencoderKLWan.from_pretrained("./model/vae", torch_dtype=torch.float16)
@@ -242,6 +242,9 @@ def track_video(n_frames,video_state):
     images = [cv2.resize(img, (W_, H_)) for img in images]
     video_state["origin_images"] = images
     images = np.array(images)
+
+    global video_predictor
+    video_predictor=video_predictor.to("cuda")
     inference_state = video_predictor.init_state(images=images/255, device=device)
     video_state["inference_state"] = inference_state