Make VAE compatible to torch.compile() (#7984)

make VAE compatible to torch.compile()
Co-authored-by: YiYi Xu <yixu310@gmail.com>

src/diffusers/models/autoencoders/autoencoder_kl.py

@@ -300,7 +300,7 @@ class AutoencoderKL(ModelMixin, ConfigMixin, FromOriginalModelMixin):
             decoded_slices = [self._decode(z_slice).sample for z_slice in z.split(1)]
             decoded = torch.cat(decoded_slices)
         else:
-            decoded = self._decode(z).sample
+            decoded = self._decode(z, return_dict=False)[0]
 
         if not return_dict:
             return (decoded,)