Mark Trainer.terminate_on_nan protected and deprecate public proper…

…ty (Lightning-AI#9849)

Co-authored-by: Adrian Wälchli <[email protected]>

CHANGELOG.md

@@ -284,6 +284,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 - Deprecated trainer argument `terminate_on_nan` in favour of `detect_anomaly`([#9175](https://github.com/PyTorchLightning/pytorch-lightning/pull/9175))
 
 
+- Deprecated `Trainer.terminate_on_nan` public attribute access ([#9849](https://github.com/PyTorchLightning/pytorch-lightning/pull/9849))
+
+
 - Deprecated `LightningModule.summarize()` in favor of `pytorch_lightning.utilities.model_summary.summarize()`
 
 

pytorch_lightning/loops/optimization/optimizer_loop.py

@@ -344,7 +344,7 @@ def backward_fn(loss: Tensor) -> None:
             self._backward(loss, optimizer, opt_idx)
 
             # check if model weights are nan
-            if self.trainer.terminate_on_nan:
+            if self.trainer._terminate_on_nan:
                 detect_nan_parameters(self.trainer.lightning_module)
 
         return backward_fn
@@ -460,7 +460,7 @@ def _training_step(self, split_batch: Any, batch_idx: int, opt_idx: int) -> Clos
 
             result = ClosureResult.from_training_step_output(training_step_output, self.trainer.accumulate_grad_batches)
 
-            if self.trainer.terminate_on_nan:
+            if self.trainer._terminate_on_nan:
                 check_finite_loss(result.closure_loss)
 
             if self.trainer.move_metrics_to_cpu:

pytorch_lightning/trainer/connectors/training_trick_connector.py

@@ -52,7 +52,7 @@ def on_trainer_init(
                 f"`track_grad_norm` should be an int, a float or 'inf' (infinity norm). Got {track_grad_norm}."
             )
 
-        self.trainer.terminate_on_nan = terminate_on_nan
+        self.trainer._terminate_on_nan = terminate_on_nan
         self.trainer.gradient_clip_val = gradient_clip_val
         self.trainer.gradient_clip_algorithm = GradClipAlgorithmType(gradient_clip_algorithm.lower())
         self.trainer.track_grad_norm = float(track_grad_norm)

pytorch_lightning/trainer/trainer.py

@@ -1990,13 +1990,6 @@ def _active_loop(self) -> Optional[Union[FitLoop, EvaluationLoop, PredictionLoop
         if self.predicting:
             return self.predict_loop
 
-    @property
-    def train_loop(self) -> FitLoop:
-        rank_zero_deprecation(
-            "`Trainer.train_loop` has been renamed to `Trainer.fit_loop` and will be removed in v1.6."
-        )
-        return self.fit_loop
-
     @property
     def _ckpt_path(self) -> Optional[str]:
         if self.state.fn == TrainerFn.VALIDATING:
@@ -2046,3 +2039,23 @@ def __getstate__(self):
 
     def __setstate__(self, state):
         self.__dict__ = state
+
+    @property
+    def train_loop(self) -> FitLoop:
+        rank_zero_deprecation(
+            "`Trainer.train_loop` has been renamed to `Trainer.fit_loop` and will be removed in v1.6."
+        )
+        return self.fit_loop
+
+    @property
+    def terminate_on_nan(self) -> bool:
+        rank_zero_deprecation("`Trainer.terminate_on_nan` is deprecated in v1.5 and will be removed in 1.7.")
+        return self._terminate_on_nan
+
+    @terminate_on_nan.setter
+    def terminate_on_nan(self, val: bool) -> None:
+        rank_zero_deprecation(
+            f"Setting `Trainer.terminate_on_nan = {val}` is deprecated in v1.5 and will be removed in 1.7."
+            f" Please set `Trainer(detect_anomaly={val})` instead."
+        )
+        self._terminate_on_nan = val  # : 212

tests/deprecated_api/test_remove_1-7.py

@@ -131,6 +131,13 @@ def test_v1_7_0_trainer_terminate_on_nan(tmpdir, terminate_on_nan):
         assert trainer.terminate_on_nan is terminate_on_nan
         assert trainer._detect_anomaly is False
 
+    trainer = Trainer()
+    with pytest.deprecated_call(match=r"`Trainer.terminate_on_nan` is deprecated in v1.5"):
+        _ = trainer.terminate_on_nan
+
+    with pytest.deprecated_call(match=r"Setting `Trainer.terminate_on_nan = True` is deprecated in v1.5"):
+        trainer.terminate_on_nan = True
+
 
 def test_v1_7_0_deprecated_on_task_dataloader(tmpdir):
     class CustomBoringModel(BoringModel):