fix: prevent second save in the end of training if last step was saved already (#36219)

* fix: prevent second save in the end of training

* fix: prevent second save in the end of training

* test: added test for no duplicate save on epoch save strategy

* fix: removed TrainerControl

* chore: style formatting

---------

Co-authored-by: JaktensTid <jaktenstid1@gmail.com>

src/transformers/trainer_callback.py

@@ -600,7 +600,7 @@ class DefaultFlowCallback(TrainerCallback):
         if state.global_step >= state.max_steps:
             control.should_training_stop = True
             # Save the model at the end if we have a save strategy
-            if args.save_strategy not in [SaveStrategy.NO, SaveStrategy.BEST]:
+            if args.save_strategy == SaveStrategy.STEPS:
                 control.should_save = True
 
         return control

tests/trainer/test_trainer_callback.py

@@ -425,3 +425,21 @@ class TrainerCallbackTest(unittest.TestCase):
         trainer.state = TrainerState.load_from_json(os.path.join(checkpoint, TRAINER_STATE_NAME))
         trainer._load_callback_state()
         assert trainer.control.should_training_stop
+
+    def test_no_duplicate_save_on_epoch_save_strategy(self):
+        times_saved = 0
+
+        class OnEndCallback(TrainerCallback):
+            def on_step_end(self, args: TrainingArguments, state: TrainerState, control, **kwargs):
+                nonlocal times_saved
+                if control.should_save:
+                    times_saved += 1
+
+            def on_epoch_end(self, args: TrainingArguments, state: TrainerState, control, **kwargs):
+                nonlocal times_saved
+                if control.should_save:
+                    times_saved += 1
+
+        trainer = self.get_trainer(max_steps=2, save_strategy="epoch", callbacks=[OnEndCallback])
+        trainer.train()
+        assert times_saved == 1