diff --git a/src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth.py b/src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth.py
index d7c1322aed..b665d4c89c 100644
--- a/src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth.py
+++ b/src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth.py
@@ -73,7 +73,7 @@ def tensor2vid(video: torch.Tensor, processor, output_type="np"):
         outputs = np.stack(outputs)
 
     if output_type == "pt":
-        outputs = torch.cat(outputs, dim=0)
+        outputs = torch.stack(outputs)
 
     return outputs