Avoid unnecessary list creation (#8595)

pytorch_lightning/accelerators/gpu.py

@@ -49,7 +49,7 @@ class GPUAccelerator(Accelerator):
     def set_nvidia_flags(local_rank: int) -> None:
         # set the correct cuda visible devices (using pci order)
         os.environ["CUDA_DEVICE_ORDER"] = "PCI_BUS_ID"
-        all_gpu_ids = ",".join([str(x) for x in range(torch.cuda.device_count())])
+        all_gpu_ids = ",".join(str(x) for x in range(torch.cuda.device_count()))
         devices = os.getenv("CUDA_VISIBLE_DEVICES", all_gpu_ids)
         _log.info(f"LOCAL_RANK: {local_rank} - CUDA_VISIBLE_DEVICES: [{devices}]")
 

pytorch_lightning/loggers/base.py

@@ -417,11 +417,11 @@ class LoggerCollection(LightningLoggerBase):
 
     @property
     def name(self) -> str:
-        return "_".join([str(logger.name) for logger in self._logger_iterable])
+        return "_".join(str(logger.name) for logger in self._logger_iterable)
 
     @property
     def version(self) -> str:
-        return "_".join([str(logger.version) for logger in self._logger_iterable])
+        return "_".join(str(logger.version) for logger in self._logger_iterable)
 
 
 class DummyExperiment:

pytorch_lightning/profiler/simple.py

@@ -88,7 +88,7 @@ class SimpleProfiler(BaseProfiler):
         if self.extended:
 
             if len(self.recorded_durations) > 0:
-                max_key = np.max([len(k) for k in self.recorded_durations.keys()])
+                max_key = max(len(k) for k in self.recorded_durations.keys())
 
                 def log_row(action, mean, num_calls, total, per):
                     row = f"{sep}{action:<{max_key}s}\t|  {mean:<15}\t|"

pytorch_lightning/utilities/debugging.py

@@ -163,7 +163,7 @@ class InternalDebugger:
 
     @property
     def num_seen_sanity_check_batches(self):
-        count = len([x for x in self.saved_val_losses if x["sanity_check"]])
+        count = sum(1 for x in self.saved_val_losses if x["sanity_check"])
         return count
 
     @property

tests/callbacks/test_stochastic_weight_avg.py

@@ -210,7 +210,7 @@ def test_trainer_and_stochastic_weight_avg(tmpdir, use_callbacks: bool, stochast
     )
     trainer.fit(model)
     if use_callbacks or stochastic_weight_avg:
-        assert len([cb for cb in trainer.callbacks if isinstance(cb, StochasticWeightAveraging)]) == 1
+        assert sum(1 for cb in trainer.callbacks if isinstance(cb, StochasticWeightAveraging)) == 1
         assert trainer.callbacks[0]._swa_lrs == (1e-3 if use_callbacks else 0.1)
     else:
         assert all(not isinstance(cb, StochasticWeightAveraging) for cb in trainer.callbacks)

tests/checkpointing/test_model_checkpoint.py

@@ -1039,7 +1039,7 @@ def test_configure_model_checkpoint(tmpdir):
 
     # default configuration
     trainer = Trainer(checkpoint_callback=True, callbacks=[], **kwargs)
-    assert len([c for c in trainer.callbacks if isinstance(c, ModelCheckpoint)]) == 1
+    assert sum(1 for c in trainer.callbacks if isinstance(c, ModelCheckpoint)) == 1
     assert isinstance(trainer.checkpoint_callback, ModelCheckpoint)
 
     # custom callback passed to callbacks list, checkpoint_callback=True is ignored

tests/trainer/test_trainer.py

@@ -556,7 +556,7 @@ def test_trainer_min_steps_and_min_epochs_not_reached(tmpdir, caplog):
         trainer.fit(model)
 
     message = f"minimum epochs ({min_epochs}) or minimum steps (None) has not been met. Training will continue"
-    num_messages = len([record.message for record in caplog.records if message in record.message])
+    num_messages = sum(1 for record in caplog.records if message in record.message)
     assert num_messages == min_epochs - 2
     assert model.training_step_invoked == min_epochs * 2