.

main.py

@@ -217,7 +217,7 @@ def predict_next_frame(previous_frames: List[np.ndarray], previous_actions: List
     data_std = 6.78
     data_min = -27.681446075439453
     data_max = 30.854148864746094
-    #image_sequence_tensor = (image_sequence_tensor - data_mean) / data_std
+    image_sequence_tensor = (image_sequence_tensor - data_mean) / data_std
     
     # Prepare the prompt based on the previous actions
     action_descriptions = []

utils.py

@@ -101,7 +101,7 @@ def sample_frame(model: LatentDiffusion, prompt: str, image_sequence: torch.Tens
             data_min = -27.681446075439453
             data_max = 30.854148864746094
             x_samples_ddim = samples_ddim
-            #x_samples_ddim = xsamples_ddim * data_std + data_mean
+            x_samples_ddim = x_samples_ddim * data_std + data_mean
             x_samples_ddim = model.decode_first_stage(x_samples_ddim)
         print ('dfsf3')
         #x_samples_ddim = pos_map.to(c['c_concat'].device).unsqueeze(0).expand(-1, 3, -1, -1)