Add warnings if training args differ from checkpoint trainer state (#29255)

* add warnings if training args differ from checkpoint args stored in trainer_state.json

* run formatting and styling

* add a test

* format and styling

---------
Co-authored-by: Jonathan Flynn <jonl.flynn@guardian.co.uk>

src/transformers/trainer.py

@@ -1529,6 +1529,29 @@ class Trainer:
 
         return model
 
+    def compare_trainer_and_checkpoint_args(self, training_args, trainer_state):
+        attributes_map = {
+            "logging_steps": "logging_steps",
+            "eval_steps": "eval_steps",
+            "save_steps": "save_steps",
+            "per_device_train_batch_size": "train_batch_size",
+        }
+
+        warnings_list = []
+        for arg_attr, state_attr in attributes_map.items():
+            arg_value = getattr(training_args, arg_attr, None)
+            state_value = getattr(trainer_state, state_attr, None)
+
+            if arg_value is not None and state_value is not None and arg_value != state_value:
+                warnings_list.append(
+                    f"Warning: The training argument '{arg_attr}' value ({arg_value}) does not match the trainer state '{state_attr}' value ({state_value}). "
+                    f"This argument will be overridden by the one found in trainer_state.json within the checkpoint directory."
+                )
+
+        if warnings_list:
+            for warning in warnings_list:
+                logger.warning(warning)
+
     def _wrap_model(self, model, training=True, dataloader=None):
         if self.args.use_ipex:
             dtype = torch.bfloat16 if self.use_cpu_amp else torch.float32
@@ -1991,6 +2014,7 @@ class Trainer:
             os.path.join(resume_from_checkpoint, TRAINER_STATE_NAME)
         ):
             self.state = TrainerState.load_from_json(os.path.join(resume_from_checkpoint, TRAINER_STATE_NAME))
+            self.compare_trainer_and_checkpoint_args(self.args, self.state)
             epochs_trained = self.state.global_step // num_update_steps_per_epoch
             if not args.ignore_data_skip:
                 steps_trained_in_current_epoch = self.state.global_step % (num_update_steps_per_epoch)

tests/trainer/test_trainer.py

@@ -2485,6 +2485,46 @@ class TrainerIntegrationTest(TestCasePlus, TrainerIntegrationCommon):
             trainer.state.best_model_checkpoint = os.path.join(tmp_dir, "checkpoint-5")
             self.check_checkpoint_deletion(trainer, tmp_dir, [5, 25])
 
+    def test_compare_trainer_and_checkpoint_args_logging(self):
+        logger = logging.get_logger()
+
+        with tempfile.TemporaryDirectory() as tmpdir, CaptureLogger(logger) as cl:
+            trainer = get_regression_trainer(
+                output_dir=tmpdir,
+                train_len=128,
+                eval_steps=5,
+                gradient_accumulation_steps=2,
+                per_device_train_batch_size=4,
+                save_steps=5,
+                learning_rate=0.1,
+            )
+            trainer.train()
+
+            checkpoint = os.path.join(tmpdir, "checkpoint-5")
+            checkpoint_trainer = get_regression_trainer(
+                output_dir=tmpdir,
+                train_len=256,
+                eval_steps=10,
+                gradient_accumulation_steps=4,
+                per_device_train_batch_size=8,
+                save_steps=10,
+                learning_rate=0.1,
+            )
+            checkpoint_trainer.train(resume_from_checkpoint=checkpoint)
+
+        self.assertIn(
+            "Warning: The training argument 'save_steps' value (10) does not match the trainer state 'save_steps' value (5). This argument will be overridden by the one found in trainer_state.json within the checkpoint directory.",
+            cl.out,
+        )
+        self.assertIn(
+            "Warning: The training argument 'per_device_train_batch_size' value (8) does not match the trainer state 'train_batch_size' value (4). This argument will be overridden by the one found in trainer_state.json within the checkpoint directory.",
+            cl.out,
+        )
+        self.assertIn(
+            "Warning: The training argument 'eval_steps' value (10) does not match the trainer state 'eval_steps' value (5). This argument will be overridden by the one found in trainer_state.json within the checkpoint directory.",
+            cl.out,
+        )
+
     def check_mem_metrics(self, trainer, check_func):
         metrics = trainer.train().metrics
         check_func("init_mem_cpu_alloc_delta", metrics)