update IP-adapter code in UNetMotionModel (huggingface#6828)

fix

Co-authored-by: yiyixuxu <yixu310@gmail,com>

Author: yiyixuxu

src/diffusers/models/unets/unet_motion_model.py

@@ -792,17 +792,17 @@ def forward(
 
         emb = self.time_embedding(t_emb, timestep_cond)
         emb = emb.repeat_interleave(repeats=num_frames, dim=0)
+        encoder_hidden_states = encoder_hidden_states.repeat_interleave(repeats=num_frames, dim=0)
 
         if self.encoder_hid_proj is not None and self.config.encoder_hid_dim_type == "ip_image_proj":
             if "image_embeds" not in added_cond_kwargs:
                 raise ValueError(
                     f"{self.__class__} has the config param `encoder_hid_dim_type` set to 'ip_image_proj' which requires the keyword argument `image_embeds` to be passed in  `added_conditions`"
                 )
             image_embeds = added_cond_kwargs.get("image_embeds")
-            image_embeds = self.encoder_hid_proj(image_embeds).to(encoder_hidden_states.dtype)
-            encoder_hidden_states = torch.cat([encoder_hidden_states, image_embeds], dim=1)
-
-        encoder_hidden_states = encoder_hidden_states.repeat_interleave(repeats=num_frames, dim=0)
+            image_embeds = self.encoder_hid_proj(image_embeds)
+            image_embeds = [image_embed.repeat_interleave(repeats=num_frames, dim=0) for image_embed in image_embeds]
+            encoder_hidden_states = (encoder_hidden_states, image_embeds)
 
         # 2. pre-process
         sample = sample.permute(0, 2, 1, 3, 4).reshape((sample.shape[0] * num_frames, -1) + sample.shape[3:])