7107 Add support to validate at training start (#7108)

Fixes #7107 .

### Description

This PR added support to optionally execute validation at training start
first, this is useful for transfer learning to validate the initial
model.

### Types of changes
<!--- Put an `x` in all the boxes that apply, and remove the not
applicable items -->
- [x] Non-breaking change (fix or new feature that would not break
existing functionality).
- [ ] Breaking change (fix or new feature that would cause existing
functionality to change).
- [ ] New tests added to cover the changes.
- [ ] Integration tests passed locally by running `./runtests.sh -f -u
--net --coverage`.
- [ ] Quick tests passed locally by running `./runtests.sh --quick
--unittests --disttests`.
- [ ] In-line docstrings updated.
- [ ] Documentation updated, tested `make html` command in the `docs/`
folder.

Signed-off-by: Nic Ma <nma@nvidia.com>

Author: Nic-Ma

monai/handlers/validation_handler.py

@@ -31,14 +31,19 @@ class ValidationHandler:
 
     """
 
-    def __init__(self, interval: int, validator: Evaluator | None = None, epoch_level: bool = True) -> None:
+    def __init__(
+        self, interval: int, validator: Evaluator | None = None, epoch_level: bool = True, exec_at_start: bool = False
+    ) -> None:
         """
         Args:
             interval: do validation every N epochs or every N iterations during training.
             validator: run the validator when trigger validation, suppose to be Evaluator.
                 if None, should call `set_validator()` before training.
             epoch_level: execute validation every N epochs or N iterations.
                 `True` is epoch level, `False` is iteration level.
+            exec_at_start: whether to execute a validation first when starting the training.
+                default to `False`. It can be useful especially for some transfer-learning cases
+                to validate the initial model before training.
 
         Raises:
             TypeError: When ``validator`` is not a ``monai.engines.evaluator.Evaluator``.
@@ -49,6 +54,7 @@ def __init__(self, interval: int, validator: Evaluator | None = None, epoch_leve
         self.validator = validator
         self.interval = interval
         self.epoch_level = epoch_level
+        self.exec_at_start = exec_at_start
 
     def set_validator(self, validator: Evaluator) -> None:
         """
@@ -67,6 +73,8 @@ def attach(self, engine: Engine) -> None:
             engine.add_event_handler(Events.EPOCH_COMPLETED(every=self.interval), self)
         else:
             engine.add_event_handler(Events.ITERATION_COMPLETED(every=self.interval), self)
+        if self.exec_at_start:
+            engine.add_event_handler(Events.STARTED, self)
 
     def __call__(self, engine: Engine) -> None:
         """

tests/test_handler_validation.py

@@ -39,8 +39,11 @@ def _train_func(engine, batch):
         # set up testing handler
         val_data_loader = torch.utils.data.DataLoader(Dataset(data))
         evaluator = TestEvaluator(torch.device("cpu:0"), val_data_loader)
-        saver = ValidationHandler(interval=2, validator=evaluator)
-        saver.attach(engine)
+        ValidationHandler(interval=2, validator=evaluator, exec_at_start=True).attach(engine)
+        # test execution at start
+        engine.run(data, max_epochs=1)
+        self.assertEqual(evaluator.state.max_epochs, 0)
+        self.assertEqual(evaluator.state.epoch_length, 8)
 
         engine.run(data, max_epochs=5)
         self.assertEqual(evaluator.state.max_epochs, 4)