fix formatting with yapf version

src/base_experiment.py

@@ -131,10 +131,9 @@ def on_validation_epoch_end(self):
             self.log(f'val/{metric}', mean, on_epoch=True, on_step=False)
         # Report the averages to ray/tune for hparam optimization.
         if self.enable_tune:
-            train.report({
-                f'val/{metric}': mean
-                for metric, mean in avgs.items()
-            })
+            train.report(
+                {f'val/{metric}': mean
+                 for metric, mean in avgs.items()})
 
     def configure_optimizers(self) -> OptimizerLRScheduler:
         optimizer = optim.Adam(self.trainable_parameters(),

src/dataset/grasp_and_lift_eeg.py

@@ -192,8 +192,8 @@ def load_from_bin(self,
 
     def download(self, root: str):
         zip_path = os.path.join(root, 'grasp-and-lift-eeg-detection.zip')
-        if not os.path.exists(zip_path) or os.path.getsize(
-                zip_path) != self.ZIP_SIZE_BYTES:
+        if not os.path.exists(
+                zip_path) or os.path.getsize(zip_path) != self.ZIP_SIZE_BYTES:
             print(f'\nDownloading from {self.ZIP_URL}')
             start = time.time()
             r = requests.get(self.ZIP_URL)

src/neural_gbuffer.py

@@ -32,10 +32,9 @@ def training_step(self, batch, batch_idx):
         recons = self.model(*labels)
         train_loss = self.model.loss_function(
             recons, orig, **self.params.get('loss_params', {}))
-        self.logger.experiment.log({
-            key: val.item()
-            for key, val in train_loss.items()
-        })
+        self.logger.experiment.log(
+            {key: val.item()
+             for key, val in train_loss.items()})
         if self.global_step > 0:
             for plot, val_indices in zip(self.plots, self.val_indices):
                 if self.global_step % plot['sample_every_n_steps'] == 0: