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

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

Author: pre-commit-ci[bot]

docs/source/solo/utils.rst

@@ -257,4 +257,3 @@ forward
 ~~~~~~~
 .. automethod:: solo.utils.positional_encoding.Summer.forward
    :noindex:
-

main_linear.py

@@ -203,9 +203,9 @@ def main(cfg: DictConfig):
             "logger": wandb_logger if cfg.wandb.enabled else None,
             "callbacks": callbacks,
             "enable_checkpointing": False,
-            "strategy": DDPStrategy(find_unused_parameters=False)
-            if cfg.strategy == "ddp"
-            else cfg.strategy,
+            "strategy": (
+                DDPStrategy(find_unused_parameters=False) if cfg.strategy == "ddp" else cfg.strategy
+            ),
         }
     )
     trainer = Trainer(**trainer_kwargs)

main_pretrain.py

@@ -221,9 +221,9 @@ def main(cfg: DictConfig):
             "logger": wandb_logger if cfg.wandb.enabled else None,
             "callbacks": callbacks,
             "enable_checkpointing": False,
-            "strategy": DDPStrategy(find_unused_parameters=False)
-            if cfg.strategy == "ddp"
-            else cfg.strategy,
+            "strategy": (
+                DDPStrategy(find_unused_parameters=False) if cfg.strategy == "ddp" else cfg.strategy
+            ),
         }
     )
     trainer = Trainer(**trainer_kwargs)

scripts/pretrain/cifar/all4one.yaml

@@ -28,7 +28,7 @@ data:
   dataset: cifar100 # change here for cifar10
   train_path: "./datasets/"
   val_path: "./datasets/"
-  format: "image_folder" 
+  format: "image_folder"
   num_workers: 4
 optimizer:
   name: "lars"

solo/data/temp_dali_fix.py

@@ -115,7 +115,5 @@ def reset(self):
                     with p._check_api_type_scope(types.PipelineAPIType.ITERATOR):
                         p.schedule_run()
         else:
-            logging.warning(
-                "DALI iterator does not support resetting while epoch is not finished. \
-                             Ignoring..."
-            )
+            logging.warning("DALI iterator does not support resetting while epoch is not finished. \
+                             Ignoring...")

solo/methods/__init__.py

@@ -39,7 +39,6 @@
 from solo.methods.wmse import WMSE
 from solo.methods.all4one import All4One
 
-
 METHODS = {
     # base classes
     "base": BaseMethod,

solo/methods/base.py

@@ -389,9 +389,11 @@ def configure_optimizers(self) -> Tuple[List, List]:
         if idxs_no_scheduler:
             partial_fn = partial(
                 static_lr,
-                get_lr=scheduler["scheduler"].get_lr
-                if isinstance(scheduler, dict)
-                else scheduler.get_lr,
+                get_lr=(
+                    scheduler["scheduler"].get_lr
+                    if isinstance(scheduler, dict)
+                    else scheduler.get_lr
+                ),
                 param_group_indexes=idxs_no_scheduler,
                 lrs_to_replace=[self.lr] * len(idxs_no_scheduler),
             )

solo/utils/misc.py

@@ -30,7 +30,6 @@
 from timm.models.helpers import group_parameters
 from timm.optim.optim_factory import _layer_map
 
-
 try:
     from solo.data.h5_dataset import H5Dataset
 except ImportError:

solo/utils/positional_encodings.py

@@ -100,9 +100,7 @@ def forward(self, tensor):
         sin_inp_y = torch.einsum("i,j->ij", pos_y, self.inv_freq)
         emb_x = get_emb(sin_inp_x).unsqueeze(1)
         emb_y = get_emb(sin_inp_y)
-        emb = torch.zeros((x, y, self.channels * 2), device=tensor.device).type(
-            tensor.type()
-        )
+        emb = torch.zeros((x, y, self.channels * 2), device=tensor.device).type(tensor.type())
         emb[:, :, : self.channels] = emb_x
         emb[:, :, self.channels : 2 * self.channels] = emb_y
 
@@ -165,9 +163,7 @@ def forward(self, tensor):
         emb_x = get_emb(sin_inp_x).unsqueeze(1).unsqueeze(1)
         emb_y = get_emb(sin_inp_y).unsqueeze(1)
         emb_z = get_emb(sin_inp_z)
-        emb = torch.zeros((x, y, z, self.channels * 3), device=tensor.device).type(
-            tensor.type()
-        )
+        emb = torch.zeros((x, y, z, self.channels * 3), device=tensor.device).type(tensor.type())
         emb[:, :, :, : self.channels] = emb_x
         emb[:, :, :, self.channels : 2 * self.channels] = emb_y
         emb[:, :, :, 2 * self.channels :] = emb_z