Fix gradient checkpointing issue for Stable Diffusion 3 (#8542)

Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
Co-authored-by: YiYi Xu <yixu310@gmail.com>

Author: 3 people

src/diffusers/models/transformers/transformer_sd3.py

@@ -306,7 +306,7 @@ def custom_forward(*inputs):
                     return custom_forward
 
                 ckpt_kwargs: Dict[str, Any] = {"use_reentrant": False} if is_torch_version(">=", "1.11.0") else {}
-                hidden_states = torch.utils.checkpoint.checkpoint(
+                encoder_hidden_states, hidden_states = torch.utils.checkpoint.checkpoint(
                     create_custom_forward(block),
                     hidden_states,
                     encoder_hidden_states,