update

app.py

@@ -173,17 +173,13 @@ def run(cfg: DictConfig):
     memory_frames.append(load_image_as_tensor(DEFAULT_IMAGE))
 
     @spaces.GPU()
-    def run_interactive(first_frame, action, first_pose, curr_frame, device):
-        global algo
+    def run_interactive(algo, first_frame, action, first_pose, curr_frame, device):
         new_frame = algo.interactive(first_frame,
                                         action,
                                         first_pose, 
                                         curr_frame, 
                                         device=device)
-        print("run????")
-        print(algo.frames)
-        print(new_frame)
-        return new_frame
+        return algo, new_frame
 
     def set_denoising_steps(denoising_steps, sampling_timesteps_state):
         algo.sampling_timesteps = denoising_steps
@@ -201,7 +197,7 @@ def run(cfg: DictConfig):
         
         for i in range(len(actions)):
             memory_curr_frame += 1
-            new_frame = run_interactive(memory_frames[0],
+            algo, new_frame = run_interactive(algo, memory_frames[0],
                                           actions[i],
                                           None, 
                                           memory_curr_frame, 
@@ -233,7 +229,7 @@ def run(cfg: DictConfig):
         memory_curr_frame = 0
         input_history = ""
 
-        _ = run_interactive(memory_frames[0],
+        algo, _ = run_interactive(algo, memory_frames[0],
                                     actions[0],
                                     poses[0],
                                     memory_curr_frame,
@@ -246,7 +242,7 @@ def run(cfg: DictConfig):
         reset()
         return SELECTED_IMAGE
 
-    _ = run_interactive(memory_frames[0],
+    algo, _ = run_interactive(algo, memory_frames[0],
                                 actions[0],
                                 poses[0],
                                 memory_curr_frame,