From bf1394a47263d828054ce3e286c622fdb82a64d1 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Adrian=20W=C3=A4lchli?= <aedu.waelchli@gmail.com>
Date: Mon, 3 May 2021 22:20:48 +0200
Subject: [PATCH] improve early stopping verbose logging (#6811)

---
 CHANGELOG.md                                  |  3 +++
 pytorch_lightning/callbacks/early_stopping.py | 26 ++++++++++++++++---
 2 files changed, 26 insertions(+), 3 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 6573a11692..f8944e2e10 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -145,6 +145,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 - Added warning when missing `Callback` and using `resume_from_checkpoint` ([#7254](https://github.com/PyTorchLightning/pytorch-lightning/pull/7254))
 
 
+- Improved verbose logging for `EarlyStopping` callback ([#6811](https://github.com/PyTorchLightning/pytorch-lightning/pull/6811))
+
+
 ### Changed
 
 
diff --git a/pytorch_lightning/callbacks/early_stopping.py b/pytorch_lightning/callbacks/early_stopping.py
index 680bc55ed2..ba9ab188d4 100644
--- a/pytorch_lightning/callbacks/early_stopping.py
+++ b/pytorch_lightning/callbacks/early_stopping.py
@@ -24,6 +24,7 @@ from typing import Any, Callable, Dict, Optional, Tuple
 import numpy as np
 import torch
 
+import pytorch_lightning as pl
 from pytorch_lightning.callbacks.base import Callback
 from pytorch_lightning.utilities import rank_zero_warn
 from pytorch_lightning.utilities.exceptions import MisconfigurationException
@@ -196,8 +197,8 @@ class EarlyStopping(Callback):
         trainer.should_stop = trainer.should_stop or should_stop
         if should_stop:
             self.stopped_epoch = trainer.current_epoch
-        if reason:
-            log.info(f"[{trainer.global_rank}] {reason}")
+        if reason and self.verbose:
+            self._log_info(trainer, reason)
 
     def _evalute_stopping_criteria(self, current: torch.Tensor) -> Tuple[bool, str]:
         should_stop = False
@@ -224,6 +225,7 @@ class EarlyStopping(Callback):
             )
         elif self.monitor_op(current - self.min_delta, self.best_score):
             should_stop = False
+            reason = self._improvement_message(current)
             self.best_score = current
             self.wait_count = 0
         else:
@@ -231,8 +233,26 @@ class EarlyStopping(Callback):
             if self.wait_count >= self.patience:
                 should_stop = True
                 reason = (
-                    f"Monitored metric {self.monitor} did not improve in the last {self.wait_count} epochs."
+                    f"Monitored metric {self.monitor} did not improve in the last {self.wait_count} records."
                     f" Best score: {self.best_score:.3f}. Signaling Trainer to stop."
                 )
 
         return should_stop, reason
+
+    def _improvement_message(self, current: torch.Tensor) -> str:
+        """ Formats a log message that informs the user about an improvement in the monitored score. """
+        if torch.isfinite(self.best_score):
+            msg = (
+                f"Metric {self.monitor} improved by {abs(self.best_score - current):.3f} >="
+                f" min_delta = {abs(self.min_delta)}. New best score: {current:.3f}"
+            )
+        else:
+            msg = f"Metric {self.monitor} improved. New best score: {current:.3f}"
+        return msg
+
+    @staticmethod
+    def _log_info(trainer: Optional["pl.Trainer"], message: str) -> None:
+        if trainer is not None and trainer.world_size > 1:
+            log.info(f"[rank: {trainer.global_rank}] {message}")
+        else:
+            log.info(message)