[pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Author: pre-commit-ci[bot]

torch_points3d/core/instantiator.py

@@ -39,7 +39,7 @@ def instantiate(self, *args, **kwargs):
 class HydraInstantiator(Instantiator):
     def litmodel(self, cfg: DictConfig, data_module: pl.LightningDataModule) -> "PointCloudBaseModel":
         return self.instantiate(cfg, instantiator=self, data_module=data_module)
-        
+
     def backbone(self, cfg: DictConfig, feature_dimension: int):
         return self.instantiate(cfg, input_nc=feature_dimension)
 

torch_points3d/models/base_model.py

@@ -18,7 +18,7 @@ def __init__(
         optimizer: OptimizerConfig,
         instantiator: Instantiator,
         data_module: pl.LightningDataModule,
-        scheduler: SchedulerConfig = None, # scheduler shouldn't be required
+        scheduler: SchedulerConfig = None,  # scheduler shouldn't be required
     ):
         super().__init__()
         # some optimizers/schedulers need parameters only known dynamically
@@ -52,7 +52,7 @@ def set_input(self, data):
             self.labels = data.y
         else:
             self.labels = None
-    
+
     def forward(self):
         features = self.backbone(self.input).x
         logits = self.head(features)
@@ -61,7 +61,7 @@ def forward(self):
         # only compute loss if loss is defined and the dset has labels
         if self.labels is not None and self.criterion is not None:
             return self.criterion(self.output, self.labels)
-            
+
     def configure_optimizers(self) -> Dict:
         """Prepare optimizer and scheduler"""
         optims = {}

torch_points3d/models/segmentation/base_model.py

@@ -4,13 +4,13 @@
 
 # Overrides PointCloudBaseModel to create segmentation-specific models
 class SegmentationBaseModel(PointCloudBaseModel):
-    
+
     # Most segmentation models will share the same head and criterion
     # Note that we don't even have to create a sparseconv3d subclass of PointCloudBaseModel
     # because the actual training logic is contained all within this and PointCloudBaseModel
     # so all we need to do is define a `backend` (in this case sparseconv3d)
     def _build_model(self, backbone_cfg):
         super()._build_model(backbone_cfg)
-        
+
         self.head = nn.Sequential(nn.Linear(self.backbone.output_nc, self.num_classes))
-        self.criterion = F.nll_loss
+        self.criterion = F.nll_loss

torch_points3d/trainer.py

@@ -30,12 +30,12 @@ def __init__(
                 "The instantiator did not return a DataModule instance." " Hydra hint: is `dataset._target_` defined?`"
             )
         self.data_module.setup("fit")
-        
+
         self.litmodel: PointCloudBaseModel = instantiator.litmodel(model, self.data_module)
         print(self.litmodel)
         self.trainer = instantiator.trainer(
             trainer,
-            logger=None, # eventually add logger config back in
+            logger=None,  # eventually add logger config back in
         )
 
     def train(self):