Add logger flag to save_hyperparameters

CHANGELOG.md

@@ -116,7 +116,10 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 - Added support `LightningModule.save_hyperparameters` when `LightningModule` is a dataclass ([#7992](https://github.com/PyTorchLightning/pytorch-lightning/pull/7992))
 
 
-- Add support for overriding `optimizer_zero_grad` and `optimizer_step` when using accumulate_grad_batches ([#7980](https://github.com/PyTorchLightning/pytorch-lightning/pull/7980))
+- Added support for overriding `optimizer_zero_grad` and `optimizer_step` when using accumulate_grad_batches ([#7980](https://github.com/PyTorchLightning/pytorch-lightning/pull/7980))
+
+
+- Added `logger` boolean flag to `save_hyperparameters` ([#7960](https://github.com/PyTorchLightning/pytorch-lightning/pull/7960))
 
 
 - Add support for calling scripts using the module syntax (`python -m package.script`) ([#8073](https://github.com/PyTorchLightning/pytorch-lightning/pull/8073))

pytorch_lightning/core/datamodule.py

@@ -97,6 +97,8 @@ def __init__(
         self._has_teardown_test = False
         self._has_teardown_predict = False
 
+        self._log_hyperparams = True
+
     @property
     def train_transforms(self):
         """

pytorch_lightning/core/lightning.py

@@ -108,6 +108,7 @@ def __init__(self, *args: Any, **kwargs: Any) -> None:
         self._automatic_optimization: bool = True
         self._truncated_bptt_steps: int = 0
         self._param_requires_grad_state = dict()
+        self._log_hyperparams = True
         self._metric_attributes: Optional[Dict[int, str]] = None
 
         # deprecated, will be removed in 1.6

pytorch_lightning/trainer/trainer.py

@@ -907,20 +907,27 @@ def _pre_dispatch(self):
 
     def _log_hyperparams(self):
         # log hyper-parameters
+        hparams_initial = None
+
         if self.logger is not None:
             # save exp to get started (this is where the first experiment logs are written)
-            datamodule_hparams = self.datamodule.hparams_initial if self.datamodule is not None else {}
-            lightning_hparams = self.lightning_module.hparams_initial
-            colliding_keys = lightning_hparams.keys() & datamodule_hparams.keys()
-            if colliding_keys:
-                raise MisconfigurationException(
-                    f"Error while merging hparams: the keys {colliding_keys} are present "
-                    "in both the LightningModule's and LightningDataModule's hparams."
-                )
-
-            hparams_initial = {**lightning_hparams, **datamodule_hparams}
-
-            self.logger.log_hyperparams(hparams_initial)
+            if self.lightning_module._log_hyperparams and self.datamodule._log_hyperparams:
+                datamodule_hparams = self.datamodule.hparams_initial if self.datamodule is not None else {}
+                lightning_hparams = self.lightning_module.hparams_initial
+                colliding_keys = lightning_hparams.keys() & datamodule_hparams.keys()
+                if colliding_keys:
+                    raise MisconfigurationException(
+                        f"Error while merging hparams: the keys {colliding_keys} are present "
+                        "in both the LightningModule's and LightningDataModule's hparams."
+                    )
+                hparams_initial = {**lightning_hparams, **datamodule_hparams}
+            elif self.lightning_module._log_hyperparams:
+                hparams_initial = self.lightning_module.hparams_initial
+            elif self.datamodule._log_hyperparams:
+                hparams_initial = self.datamodule.hparams_initial
+
+            if hparams_initial is not None:
+                self.logger.log_hyperparams(hparams_initial)
             self.logger.log_graph(self.lightning_module)
             self.logger.save()
 

pytorch_lightning/utilities/hparams_mixin.py

@@ -30,7 +30,8 @@ def save_hyperparameters(
         self,
         *args,
         ignore: Optional[Union[Sequence[str], str]] = None,
-        frame: Optional[types.FrameType] = None
+        frame: Optional[types.FrameType] = None,
+        logger: bool = True
     ) -> None:
         """Save arguments to ``hparams`` attribute.
 
@@ -40,6 +41,7 @@ def save_hyperparameters(
             ignore: an argument name or a list of argument names from
                 class ``__init__`` to be ignored
             frame: a frame object. Default is None
+            logger: Whether to save hyperparameters by logger. Default: True
 
         Example::
             >>> class ManuallyArgsModel(HyperparametersMixin):
@@ -92,6 +94,7 @@ class ``__init__`` to be ignored
             "arg1": 1
             "arg3": 3.14
         """
+        self._log_hyperparams = logger
         # the frame needs to be created in this file.
         if not frame:
             frame = inspect.currentframe().f_back

tests/loggers/test_base.py

@@ -17,6 +17,7 @@
 from unittest.mock import MagicMock
 
 import numpy as np
+import pytest
 
 from pytorch_lightning import Trainer
 from pytorch_lightning.loggers import LightningLoggerBase, LoggerCollection, TensorBoardLogger
@@ -290,3 +291,37 @@ def log_hyperparams(self, params):
     }
     logger.log_hyperparams(Namespace(**np_params))
     assert logger.logged_params == sanitized_params
+
+
+@pytest.mark.parametrize("logger", [True, False])
+def test_log_hyperparams_being_called(tmpdir, logger):
+
+    class TestLogger(DummyLogger):
+
+        def __init__(self):
+            super().__init__()
+            self.log_hyperparams_called = False
+
+        def log_hyperparams(self, *args, **kwargs):
+            self.log_hyperparams_called = True
+
+    class TestModel(BoringModel):
+
+        def __init__(self, param_one, param_two):
+            super().__init__()
+            self.save_hyperparameters(logger=logger)
+
+    test_logger = TestLogger()
+    model = TestModel("pytorch", "lightning")
+
+    trainer = Trainer(
+        default_root_dir=tmpdir,
+        logger=test_logger,
+        max_epochs=1,
+        limit_train_batches=0.1,
+        limit_val_batches=0.1,
+        num_sanity_val_steps=0,
+    )
+    trainer.fit(model)
+
+    assert logger == test_logger.log_hyperparams_called