[Fix] No training log when the num of iterations is smaller than the default interval

mmengine/hooks/logger_hook.py

@@ -167,7 +167,8 @@ def after_train_iter(self,
             tag, log_str = runner.log_processor.get_log_after_iter(
                 runner, batch_idx, 'train')
         elif (self.end_of_epoch(runner.train_dataloader, batch_idx)
-              and not self.ignore_last):
+              and (not self.ignore_last
+                   or len(runner.train_dataloader) <= self.interval)):
             # `runner.max_iters` may not be divisible by `self.interval`. if
             # `self.ignore_last==True`, the log of remaining iterations will
             # be recorded (Epoch [4][1000/1007], the logs of 998-1007

tests/test_hooks/test_logger_hook.py

@@ -127,6 +127,16 @@ def test_after_train_iter(self):
         logger_hook.after_train_iter(runner, batch_idx=999)
         runner.logger.info.assert_called()
 
+        # Test print training log when the num of
+        # iterations is smaller than the default interval
+        runner = MagicMock()
+        runner.log_processor.get_log_after_iter = MagicMock(
+            return_value=(dict(), 'log_str'))
+        runner.train_dataloader = [0] * 9
+        logger_hook = LoggerHook()
+        logger_hook.after_train_iter(runner, batch_idx=8)
+        runner.log_processor.get_log_after_iter.assert_called()
+
     def test_after_val_epoch(self):
         logger_hook = LoggerHook()
         runner = MagicMock()

tests/test_runner/test_runner.py

@@ -1727,6 +1727,17 @@ def train_step(self, *args, **kwargs):
         with self.assertRaisesRegex(AssertionError, 'If you want to validate'):
             runner.train()
 
+        # 13 Test the logs will be printed when the length of
+        # train_dataloader is smaller than the interval set in LoggerHook
+        cfg = copy.deepcopy(self.epoch_based_cfg)
+        cfg.experiment_name = 'test_train13'
+        cfg.default_hooks = dict(logger=dict(type='LoggerHook', interval=5))
+        runner = Runner.from_cfg(cfg)
+        runner.train()
+        with open(runner.logger._log_file) as f:
+            log = f.read()
+        self.assertIn('Epoch(train) [1][4/4]', log)
+
     @skipIf(
         SKIP_TEST_COMPILE,
         reason='torch.compile is not valid, please install PyTorch>=2.0.0')