fix

Author: sayakpaul

src/diffusers/pipelines/ovis_image/pipeline_ovis_image.py

@@ -53,6 +53,7 @@
 """
 
 
+# Copied from diffusers.pipelines.flux.pipeline_flux_utils.calculate_shift
 def calculate_shift(
     image_seq_len,
     base_seq_len: int = 256,
@@ -66,7 +67,7 @@ def calculate_shift(
     return mu
 
 
-# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.retrieve_timesteps
+# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_utils.retrieve_timesteps
 def retrieve_timesteps(
     scheduler,
     num_inference_steps: Optional[int] = None,

src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl.py

@@ -309,6 +309,24 @@ def prepare_latents(self, batch_size, num_channels_latents, height, width, dtype
         latents = latents * self.scheduler.init_noise_sigma
         return latents
 
+    def _get_add_time_ids(
+        self, original_size, crops_coords_top_left, target_size, dtype, text_encoder_projection_dim=None
+    ):
+        add_time_ids = list(original_size + crops_coords_top_left + target_size)
+
+        passed_add_embed_dim = (
+            self.unet.config.addition_time_embed_dim * len(add_time_ids) + text_encoder_projection_dim
+        )
+        expected_add_embed_dim = self.unet.add_embedding.linear_1.in_features
+
+        if expected_add_embed_dim != passed_add_embed_dim:
+            raise ValueError(
+                f"Model expects an added time embedding vector of length {expected_add_embed_dim}, but a vector of {passed_add_embed_dim} was created. The model has an incorrect config. Please check `unet.config.time_embedding_type` and `text_encoder_2.config.projection_dim`."
+            )
+
+        add_time_ids = torch.tensor([add_time_ids], dtype=dtype)
+        return add_time_ids
+
     @torch.no_grad()
     @replace_example_docstring(EXAMPLE_DOC_STRING)
     def __call__(