Fix trainer saving safetensors: metadata is None (#28219)

* Update trainer.py

* format

src/transformers/trainer.py

@@ -2908,7 +2908,9 @@ class Trainer:
             else:
                 logger.info("Trainer.model is not a `PreTrainedModel`, only saving its state dict.")
                 if self.args.save_safetensors:
-                    safetensors.torch.save_file(state_dict, os.path.join(output_dir, SAFE_WEIGHTS_NAME))
+                    safetensors.torch.save_file(
+                        state_dict, os.path.join(output_dir, SAFE_WEIGHTS_NAME), metadata={"format": "pt"}
+                    )
                 else:
                     torch.save(state_dict, os.path.join(output_dir, WEIGHTS_NAME))
         else: