Fix #4375: Always use trainer.global_step for step (#4376)

* Fix #4375: Always use trainer.global_step for step * Changelog * Remove superflous use "epoch" * Update Changelog Co-authored-by: Nicki Skafte <skaftenicki@gmail.com>
Lightning-AI · Nov 22, 2020 · 8601268 · 8601268
1 parent 299de5d
commit 8601268
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 1 deletion.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -49,6 +49,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 
 ### Changed
 
+- Consistently use `step=trainer.global_step` in `LearningRateMonitor` independently of `logging_interval` ([#4376](https://github.com/PyTorchLightning/pytorch-lightning/pull/4376))
+
+
 - Tuner algorithms will be skipped if `fast_dev_run=True` ([#3903](https://github.com/PyTorchLightning/pytorch-lightning/pull/3903))
 
 

diff --git a/pytorch_lightning/callbacks/lr_monitor.py b/pytorch_lightning/callbacks/lr_monitor.py
@@ -114,7 +114,7 @@ def on_train_epoch_start(self, trainer, *args, **kwargs):
  latest_stat = self._extract_stats(trainer, interval)
 
  if trainer.logger is not None and latest_stat:
- trainer.logger.log_metrics(latest_stat, step=trainer.current_epoch)
+ trainer.logger.log_metrics(latest_stat, step=trainer.global_step)
 
  def _extract_stats(self, trainer, interval: str) -> Dict[str, float]:
  latest_stat = {}