update

src/diffusers/pipelines/cogvideo/pipeline_cogvideox.py

@@ -429,7 +429,7 @@ def __call__(
         negative_prompt: Optional[Union[str, List[str]]] = None,
         height: int = 480,
         width: int = 720,
-        num_frames: int = 48,
+        num_frames: int = 49,
         num_inference_steps: int = 50,
         timesteps: Optional[List[int]] = None,
         guidance_scale: float = 6,
@@ -524,9 +524,9 @@ def __call__(
             `tuple`. When returning a tuple, the first element is a list with the generated images.
         """
 
-        if num_frames > 48:
+        if num_frames > 49:
             raise ValueError(
-                "The number of frames must be less than 48 for now due to static positional embeddings. This will be updated in the future to remove this limitation."
+                "The number of frames must be less than 49 for now due to static positional embeddings. This will be updated in the future to remove this limitation."
             )
 
         if isinstance(callback_on_step_end, (PipelineCallback, MultiPipelineCallbacks)):

tests/pipelines/cogvideox/test_cogvideox.py

@@ -148,8 +148,8 @@ def test_inference(self):
         video = pipe(**inputs).frames
         generated_video = video[0]
 
-        self.assertEqual(generated_video.shape, (9, 3, 16, 16))
-        expected_video = torch.randn(9, 3, 16, 16)
+        self.assertEqual(generated_video.shape, (8, 3, 16, 16))
+        expected_video = torch.randn(8, 3, 16, 16)
         max_diff = np.abs(generated_video - expected_video).max()
         self.assertLessEqual(max_diff, 1e10)