Deprecate terminate_on_nan Trainer argument in favor of detect_anomaly (#9175)

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

Author: 3 people

CHANGELOG.md

@@ -271,6 +271,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 
 ### Deprecated
 
+- Deprecated trainer argument `terminate_on_nan` in favour of `detect_anomaly`([#9175](https://github.com/PyTorchLightning/pytorch-lightning/pull/9175))
+
+
 - Deprecated `LightningModule.summarize()` in favor of `pytorch_lightning.utilities.model_summary.summarize()`
 
 

pytorch_lightning/trainer/connectors/training_trick_connector.py

@@ -11,9 +11,9 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-from typing import Union
+from typing import Optional, Union
 
-from pytorch_lightning.utilities import GradClipAlgorithmType
+from pytorch_lightning.utilities import GradClipAlgorithmType, rank_zero_deprecation
 from pytorch_lightning.utilities.exceptions import MisconfigurationException
 
 
@@ -26,10 +26,15 @@ def on_trainer_init(
         gradient_clip_val: Union[int, float],
         gradient_clip_algorithm: str,
         track_grad_norm: Union[int, float, str],
-        terminate_on_nan: bool,
+        terminate_on_nan: Optional[bool],
     ):
-        if not isinstance(terminate_on_nan, bool):
-            raise TypeError(f"`terminate_on_nan` should be a bool, got {terminate_on_nan}.")
+        if terminate_on_nan is not None:
+            rank_zero_deprecation(
+                "Trainer argument `terminate_on_nan` was deprecated in v1.5 and will be removed in 1.7."
+                " Please use `Trainer(detect_anomaly=True)` instead."
+            )
+            if not isinstance(terminate_on_nan, bool):
+                raise TypeError(f"`terminate_on_nan` should be a bool, got {terminate_on_nan}.")
 
         # gradient clipping
         if not isinstance(gradient_clip_val, (int, float)):

pytorch_lightning/trainer/trainer.py

@@ -167,7 +167,7 @@ def __init__(
         reload_dataloaders_every_epoch: bool = False,
         auto_lr_find: Union[bool, str] = False,
         replace_sampler_ddp: bool = True,
-        terminate_on_nan: bool = False,
+        detect_anomaly: bool = False,
         auto_scale_batch_size: Union[str, bool] = False,
         prepare_data_per_node: Optional[bool] = None,
         plugins: Optional[Union[PLUGIN_INPUT, List[PLUGIN_INPUT]]] = None,
@@ -177,7 +177,7 @@ def __init__(
         move_metrics_to_cpu: bool = False,
         multiple_trainloader_mode: str = "max_size_cycle",
         stochastic_weight_avg: bool = False,
-        detect_anomaly: bool = False,
+        terminate_on_nan: Optional[bool] = None,
     ):
         r"""
         Customize every aspect of training via flags.
@@ -351,6 +351,12 @@ def __init__(
             terminate_on_nan: If set to True, will terminate training (by raising a `ValueError`) at the
                 end of each training batch, if any of the parameters or the loss are NaN or +/-inf.
 
+                .. deprecated:: v1.5
+                    Trainer argument ``terminate_on_nan`` was deprecated in v1.5 and will be removed in 1.7.
+                    Please use ``detect_anomaly`` instead.
+
+            detect_anomaly: Enable anomaly detection for the autograd engine.
+
             tpu_cores: How many TPU cores to train on (1 or 8) / Single TPU to train on [1]
 
             ipus: How many IPUs to train on.

tests/deprecated_api/test_remove_1-7.py

@@ -122,6 +122,16 @@ def test_v1_7_0_stochastic_weight_avg_trainer_constructor(tmpdir):
         _ = Trainer(stochastic_weight_avg=True)
 
 
+@pytest.mark.parametrize("terminate_on_nan", [True, False])
+def test_v1_7_0_trainer_terminate_on_nan(tmpdir, terminate_on_nan):
+    with pytest.deprecated_call(
+        match="Trainer argument `terminate_on_nan` was deprecated in v1.5 and will be removed in 1.7"
+    ):
+        trainer = Trainer(terminate_on_nan=terminate_on_nan)
+        assert trainer.terminate_on_nan is terminate_on_nan
+        assert trainer._detect_anomaly is False
+
+
 def test_v1_7_0_deprecated_on_task_dataloader(tmpdir):
     class CustomBoringModel(BoringModel):
         def on_train_dataloader(self):