Ignore step param in Neptune logger's log_metric method (#5510)

* Ignore `step` param in Neptune logger's log_metric method

The `step` parameter is ignored because Neptune requires strictly increasing step values, a condition which is sometimes violated in Lighting e.g. when `fit()` and `test()` are called one after another on some models. `step` could be enabled again once Lightning guarantees that step values are always strictly increasing.

Also a minor bugfix: the `log_text()` method should use Neptune's `log_text()` method.

* Update neptune.py

* Update test_neptune.py

* Update test_all.py

* fix neptune tests

* add chlog

Co-authored-by: Adrian Wälchli <[email protected]>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Co-authored-by: Nicki Skafte <[email protected]>
(cherry picked from commit 5d76b31)

Author: 4 people

CHANGELOG.md

@@ -128,6 +128,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 - Increased TPU check timeout from 20s to 100s ([#5598](https://github.com/PyTorchLightning/pytorch-lightning/pull/5598))
 
 
+- Ignored `step` param in Neptune logger's log_metric method ([#5510](https://github.com/PyTorchLightning/pytorch-lightning/pull/5510))
+
+
 ### Deprecated
 
 - Function `stat_scores_multiple_classes` is deprecated in favor of `stat_scores` ([#4839](https://github.com/PyTorchLightning/pytorch-lightning/pull/4839))

pytorch_lightning/loggers/neptune.py

@@ -251,13 +251,15 @@ def log_metrics(
 
         Args:
             metrics: Dictionary with metric names as keys and measured quantities as values
-            step: Step number at which the metrics should be recorded, must be strictly increasing
+            step: Step number at which the metrics should be recorded, currently ignored
         """
         assert rank_zero_only.rank == 0, 'experiment tried to log from global_rank != 0'
 
         metrics = self._add_prefix(metrics)
         for key, val in metrics.items():
-            self.log_metric(key, val, step=step)
+            # `step` is ignored because Neptune expects strictly increasing step values which
+            # Lighting does not always guarantee.
+            self.log_metric(key, val)
 
     @rank_zero_only
     def finalize(self, status: str) -> None:
@@ -317,7 +319,7 @@ def log_text(self, log_name: str, text: str, step: Optional[int] = None) -> None
             text: The value of the log (data-point).
             step: Step number at which the metrics should be recorded, must be strictly increasing
         """
-        self.log_metric(log_name, text, step=step)
+        self.experiment.log_text(log_name, text, step=step)
 
     @rank_zero_only
     def log_image(self,

tests/loggers/test_all.py

@@ -373,7 +373,7 @@ def test_logger_with_prefix_all(tmpdir, monkeypatch):
     with mock.patch('pytorch_lightning.loggers.neptune.neptune'):
         logger = _instantiate_logger(NeptuneLogger, save_idr=tmpdir, prefix=prefix)
         logger.log_metrics({"test": 1.0}, step=0)
-        logger.experiment.log_metric.assert_called_once_with("tmp-test", x=0, y=1.0)
+        logger.experiment.log_metric.assert_called_once_with("tmp-test", 1.0)
 
     # TensorBoard
     with mock.patch('pytorch_lightning.loggers.tensorboard.SummaryWriter'):

tests/loggers/test_neptune.py

@@ -76,8 +76,8 @@ def test_neptune_additional_methods(neptune):
     created_experiment.log_metric.reset_mock()
 
     logger.log_text('test', 'text')
-    created_experiment.log_metric.assert_called_once_with('test', 'text')
-    created_experiment.log_metric.reset_mock()
+    created_experiment.log_text.assert_called_once_with('test', 'text', step=None)
+    created_experiment.log_text.reset_mock()
 
     logger.log_image('test', 'image file')
     created_experiment.log_image.assert_called_once_with('test', 'image file')