Update app.py

app.py

@@ -257,7 +257,10 @@ def load_model(checkpoint):
         efficienttam_checkpoint = "./checkpoints/demo/efficienttam_s.pt"
         model_cfg = "efficienttam_s.yaml"
         return [efficienttam_checkpoint, model_cfg]
-    
+
+@spaces.GPU
+@torch.inference_mode()
+@torch.autocast(device_type="cuda", dtype=torch.bfloat16)
 def get_mask_sam_process(
     stored_inference_state,
     input_first_frame_image, 
@@ -349,6 +352,9 @@ def get_mask_sam_process(
     
     return gr.update(visible=True), "output_first_frame.jpg", frame_names, predictor, inference_state, gr.update(choices=available_frames_to_check, value=working_frame, visible=True)
 
+@spaces.GPU
+@torch.inference_mode()
+@torch.autocast(device_type="cuda", dtype=torch.bfloat16)
 def propagate_to_all(tracking_points, video_in, checkpoint, stored_inference_state, stored_frame_names, video_frames_dir, vis_frame_type, available_frames_to_check, working_frame):
     if tracking_points is None or video_in is None or checkpoint is None or stored_inference_state is None:
         return gr.update(value=None), gr.update(value=None), gr.update(value=None), available_frames_to_check, gr.update(visible=False)