Raise error when using save_only_model with `load_best_model_at_end…

…` for DeepSpeed/FSDP (#28866)

* Raise error when using `save_only_model` with `load_best_model_at_end` for DeepSpeed/FSDP

* Update trainer.py

src/transformers/trainer.py

@@ -4054,6 +4054,15 @@ def create_accelerator_and_postprocess(self):
         if self.is_deepspeed_enabled and getattr(self.args, "hf_deepspeed_config", None) is None:
             self.propagate_args_to_deepspeed()
 
+        # `save_only_model` can't be used with DeepSpeed/FSDP along with `load_best_model_at_end`
+        if (
+            self.args.save_only_model
+            and (self.is_deepspeed_enabled or self.is_fsdp_enabled)
+            and self.args.load_best_model_at_end
+        ):
+            wrapper = "DeepSpeed" if self.is_deepspeed_enabled else "FSDP"
+            raise ValueError(f"{wrapper} can't be used with `save_only_model` along with `load_best_model_at_end`.")
+
     def propagate_args_to_deepspeed(self, auto_find_batch_size=False):
         """
         Sets values in the deepspeed plugin based on the Trainer args