diff --git a/d2go/checkpoint/fsdp_checkpoint.py b/d2go/checkpoint/fsdp_checkpoint.py
index 5c1f1f2b..2c33c888 100644
--- a/d2go/checkpoint/fsdp_checkpoint.py
+++ b/d2go/checkpoint/fsdp_checkpoint.py
@@ -222,9 +222,13 @@ def _save_file(self, data, filename):
     def _load_file(self, f: str):
         if isinstance(self.model, FSDPWrapper):
             with (
-                interleave_by_rank(concurrency_limit=self._concurrency_limit_fetcher())
-                if self.model.state_dict_type != StateDictType.FULL_STATE_DICT
-                else nullcontext()  # FULL_STATE_DICT doesn't need interleaving
+                (
+                    interleave_by_rank(
+                        concurrency_limit=self._concurrency_limit_fetcher()
+                    )
+                    if self.model.state_dict_type != StateDictType.FULL_STATE_DICT
+                    else nullcontext()
+                )  # FULL_STATE_DICT doesn't need interleaving
             ):
                 # use mmap for FSDP checkpoints
                 return torch.load(f, map_location=torch.device("cpu"), mmap=True)
diff --git a/d2go/config/utils.py b/d2go/config/utils.py
index 406cf6ae..610d1d21 100644
--- a/d2go/config/utils.py
+++ b/d2go/config/utils.py
@@ -209,7 +209,6 @@ def get_diff_cfg_rec(old_cfg, new_cfg, out):
                     out[key] = out.__class__()
                     out[key] = get_diff_cfg_rec(old_cfg[key], new_cfg[key], out[key])
                 else:
-
                     out[key] = new_cfg[key]
         return out
 
diff --git a/d2go/data/disk_cache.py b/d2go/data/disk_cache.py
index 9ff02e4d..1a96bed3 100644
--- a/d2go/data/disk_cache.py
+++ b/d2go/data/disk_cache.py
@@ -80,7 +80,6 @@ def _initialize_diskcache(self):
         # self._cache.cache.clear(retry=True)  # seems faster if index exists
 
         if comm.get_local_rank() == 0:
-
             if self._diskcache_strategy == "naive":
                 for i, item in enumerate(self._lst):
                     ret = self._write_to_local_db((i, item))
diff --git a/d2go/data/extended_coco.py b/d2go/data/extended_coco.py
index d017e59d..ba722e95 100644
--- a/d2go/data/extended_coco.py
+++ b/d2go/data/extended_coco.py
@@ -69,7 +69,6 @@ def extract_archive_file(archive_fn: str, im_dir: str):
 
 
 def save_converted_json(target_json, convert_coco_dict):
-
     if target_json in COCOTEXT_DATASET_CONVERSION_STATUS:
         return
 
@@ -339,7 +338,6 @@ def convert_to_dict_list(
     filter_empty_annotations: Optional[bool] = True,
     _post_process_: Optional[Callable] = None,
 ) -> List[Dict]:
-
     ann_error_report = {
         name: ErrorEntry(name, msg, 0)
         for name, msg in [
diff --git a/d2go/modeling/backbone/fbnet_v2.py b/d2go/modeling/backbone/fbnet_v2.py
index fcfb50b9..4060e5e1 100644
--- a/d2go/modeling/backbone/fbnet_v2.py
+++ b/d2go/modeling/backbone/fbnet_v2.py
@@ -59,8 +59,8 @@ def _merge_fbnetv2_arch_def(cfg):
 def _parse_arch_def(cfg):
     arch = cfg.MODEL.FBNET_V2.ARCH
     arch_def = cfg.MODEL.FBNET_V2.ARCH_DEF
-    assert (arch != "" and not arch_def) ^ (
-        not arch and arch_def != []
+    assert (
+        (arch != "" and not arch_def) ^ (not arch and arch_def != [])
     ), "Only allow one unset node between MODEL.FBNET_V2.ARCH ({}) and MODEL.FBNET_V2.ARCH_DEF ({})".format(
         arch, arch_def
     )
diff --git a/d2go/modeling/meta_arch/rcnn.py b/d2go/modeling/meta_arch/rcnn.py
index a50ad144..a4939fb6 100644
--- a/d2go/modeling/meta_arch/rcnn.py
+++ b/d2go/modeling/meta_arch/rcnn.py
@@ -132,7 +132,6 @@ def default_rcnn_prepare_for_export(self, cfg, inputs, predictor_type):
 
 
 def _apply_eager_mode_quant(cfg, model):
-
     if isinstance(model, GeneralizedRCNN):
         """Wrap each quantized part of the model to insert Quant and DeQuant in-place"""
 
diff --git a/d2go/modeling/subclass.py b/d2go/modeling/subclass.py
index 2fcae567..bb863a99 100644
--- a/d2go/modeling/subclass.py
+++ b/d2go/modeling/subclass.py
@@ -30,9 +30,7 @@ def add_subclass_configs(cfg):
     _C.MODEL.SUBCLASS.NUM_SUBCLASSES = 0  # must be set
     _C.MODEL.SUBCLASS.NUM_LAYERS = 1
     _C.MODEL.SUBCLASS.SUBCLASS_ID_FETCHER = "SubclassFetcher"  # ABC, must be set
-    _C.MODEL.SUBCLASS.SUBCLASS_MAPPING = (
-        []
-    )  # subclass mapping from model output to annotation
+    _C.MODEL.SUBCLASS.SUBCLASS_MAPPING = []  # subclass mapping from model output to annotation
 
 
 class SubclassFetcher(ABC):
diff --git a/d2go/registry/bootstrap.py b/d2go/registry/bootstrap.py
index 6c5cb666..e0c3905c 100644
--- a/d2go/registry/bootstrap.py
+++ b/d2go/registry/bootstrap.py
@@ -87,7 +87,6 @@ def _match_any(name, module_full_names, match_submodule=False):
 
 
 def _import_mock(name, globals=None, locals=None, fromlist=(), level=0):
-
     use_orig_import = False
 
     # enable some first-party packages
diff --git a/d2go/trainer/fsdp.py b/d2go/trainer/fsdp.py
index d3a46e8e..035100fd 100644
--- a/d2go/trainer/fsdp.py
+++ b/d2go/trainer/fsdp.py
@@ -130,7 +130,6 @@ def postorder_fsdp_wrap(
         parent_module: Optional[nn.Module],
         ignore_branch: bool,
     ):
-
         rank = dist.get_rank()
         # don't traverse branches of specified ignored modules
         if module in modules_not_to_fsdp:
diff --git a/d2go/utils/logging.py b/d2go/utils/logging.py
index c446da4d..ca4e9bc5 100644
--- a/d2go/utils/logging.py
+++ b/d2go/utils/logging.py
@@ -69,13 +69,11 @@ def _log_exit(category: str, name: str, unique_id: str, duration: float) -> None
 def log_interval(
     category: Optional[str] = None, name: Optional[str] = None
 ) -> Callable[[Callable[..., _T]], Callable[..., _T]]:
-
     _unique_id = uuid.uuid1().int >> 97
     _overwrite_category = category
     _overwrite_name = name
 
     def log_interval_deco(func: Callable[..., _T]) -> Callable[..., _T]:
-
         _category = _overwrite_category or func.__qualname__.split(".")[0]
         _name = _overwrite_name or func.__name__
 
diff --git a/projects_oss/detr/detr/backbone/pit.py b/projects_oss/detr/detr/backbone/pit.py
index dd4081c6..d7cce5cc 100644
--- a/projects_oss/detr/detr/backbone/pit.py
+++ b/projects_oss/detr/detr/backbone/pit.py
@@ -97,7 +97,6 @@ def __init__(
         self.fc = nn.Linear(in_feature, out_feature)
 
     def forward(self, x, cls_token):
-
         x = self.conv(x)
         cls_token = self.fc(cls_token)
 
diff --git a/projects_oss/detr/detr/datasets/coco.py b/projects_oss/detr/detr/datasets/coco.py
index 43622437..7bfc4ec6 100644
--- a/projects_oss/detr/detr/datasets/coco.py
+++ b/projects_oss/detr/detr/datasets/coco.py
@@ -6,6 +6,7 @@
 
 Mostly copy-paste from https://github.com/pytorch/vision/blob/13b35ff/references/detection/coco_utils.py
 """
+
 import os
 from pathlib import Path
 
@@ -126,7 +127,6 @@ def __call__(self, image, target):
 
 
 def make_coco_transforms(image_set):
-
     normalize = T.Compose(
         [T.ToTensor(), T.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])]
     )
diff --git a/projects_oss/detr/detr/datasets/coco_eval.py b/projects_oss/detr/detr/datasets/coco_eval.py
index adfe11b2..c8c5e91a 100644
--- a/projects_oss/detr/detr/datasets/coco_eval.py
+++ b/projects_oss/detr/detr/datasets/coco_eval.py
@@ -8,6 +8,7 @@
 The difference is that there is less copy-pasting from pycocotools
 in the end of the file, as python3 can suppress prints with contextlib
 """
+
 import contextlib
 import copy
 import os
diff --git a/projects_oss/detr/detr/datasets/transforms.py b/projects_oss/detr/detr/datasets/transforms.py
index c3dc44dc..a87d5366 100644
--- a/projects_oss/detr/detr/datasets/transforms.py
+++ b/projects_oss/detr/detr/datasets/transforms.py
@@ -4,6 +4,7 @@
 """
 Transforms and data augmentation for both image + bbox.
 """
+
 import random
 
 import PIL
diff --git a/projects_oss/detr/detr/engine.py b/projects_oss/detr/detr/engine.py
index ca771f4c..180257db 100644
--- a/projects_oss/detr/detr/engine.py
+++ b/projects_oss/detr/detr/engine.py
@@ -2,6 +2,7 @@
 """
 Train and eval functions used in main.py
 """
+
 import math
 import os
 import sys
diff --git a/projects_oss/detr/detr/functions/ms_deform_attn_func.py b/projects_oss/detr/detr/functions/ms_deform_attn_func.py
index 40862a5e..73c27509 100644
--- a/projects_oss/detr/detr/functions/ms_deform_attn_func.py
+++ b/projects_oss/detr/detr/functions/ms_deform_attn_func.py
@@ -18,7 +18,6 @@
 
 
 class MSDeformAttnFunction(Function):
-
     # The @custom_fwd and @custom_bwd decorators are used in this case to allow enabling of
     # Automatic Mixed Precision when we do not have implementations of custom CUDA kernels for
     # all the precision types.
diff --git a/projects_oss/detr/detr/models/backbone.py b/projects_oss/detr/detr/models/backbone.py
index f6bebac4..6a99e0a4 100644
--- a/projects_oss/detr/detr/models/backbone.py
+++ b/projects_oss/detr/detr/models/backbone.py
@@ -12,6 +12,7 @@
 """
 Backbone modules.
 """
+
 from collections import OrderedDict
 from typing import Dict, List
 
diff --git a/projects_oss/detr/detr/models/deformable_detr.py b/projects_oss/detr/detr/models/deformable_detr.py
index e8d312f6..d345fae4 100644
--- a/projects_oss/detr/detr/models/deformable_detr.py
+++ b/projects_oss/detr/detr/models/deformable_detr.py
@@ -10,6 +10,7 @@
 """
 Deformable DETR model and criterion classes.
 """
+
 import copy
 import math
 
diff --git a/projects_oss/detr/detr/models/detr.py b/projects_oss/detr/detr/models/detr.py
index 91103df3..9df0e4f7 100644
--- a/projects_oss/detr/detr/models/detr.py
+++ b/projects_oss/detr/detr/models/detr.py
@@ -4,6 +4,7 @@
 """
 DETR model and criterion classes.
 """
+
 import torch
 import torch.nn.functional as F
 from detectron2.config import configurable
diff --git a/projects_oss/detr/detr/models/matcher.py b/projects_oss/detr/detr/models/matcher.py
index ff2040b8..15d2a160 100644
--- a/projects_oss/detr/detr/models/matcher.py
+++ b/projects_oss/detr/detr/models/matcher.py
@@ -4,6 +4,7 @@
 """
 Modules to compute the matching cost and solve the corresponding LSAP.
 """
+
 import torch
 from detr.util.box_ops import box_cxcywh_to_xyxy, generalized_box_iou
 from scipy.optimize import linear_sum_assignment
diff --git a/projects_oss/detr/detr/models/position_encoding.py b/projects_oss/detr/detr/models/position_encoding.py
index e122c855..10bae9c3 100644
--- a/projects_oss/detr/detr/models/position_encoding.py
+++ b/projects_oss/detr/detr/models/position_encoding.py
@@ -4,6 +4,7 @@
 """
 Various positional encodings for the transformer.
 """
+
 import math
 
 import torch
@@ -63,14 +64,10 @@ def forward(self, tensor_list: NestedTensor):
         pos_y = y_embed[:, :, :, None] / dim_t
         pos_x = torch.stack(
             (pos_x[:, :, :, 0::2].sin(), pos_x[:, :, :, 1::2].cos()), dim=4
-        ).flatten(
-            3
-        )  # shape (B, H, W, N)
+        ).flatten(3)  # shape (B, H, W, N)
         pos_y = torch.stack(
             (pos_y[:, :, :, 0::2].sin(), pos_y[:, :, :, 1::2].cos()), dim=4
-        ).flatten(
-            3
-        )  # shape (B, H, W, N)
+        ).flatten(3)  # shape (B, H, W, N)
         pos = torch.cat((pos_y, pos_x), dim=3).permute(
             0, 3, 1, 2
         )  # shape (B, 2*N, H, W)
diff --git a/projects_oss/detr/detr/models/segmentation.py b/projects_oss/detr/detr/models/segmentation.py
index a8f48fad..9ffedd80 100644
--- a/projects_oss/detr/detr/models/segmentation.py
+++ b/projects_oss/detr/detr/models/segmentation.py
@@ -4,6 +4,7 @@
 """
 This file provides the definition of the convolutional heads used to predict masks, as well as the losses
 """
+
 import io
 from collections import defaultdict
 from typing import List, Optional
diff --git a/projects_oss/detr/detr/models/setcriterion.py b/projects_oss/detr/detr/models/setcriterion.py
index 221dfffe..386c2c8f 100644
--- a/projects_oss/detr/detr/models/setcriterion.py
+++ b/projects_oss/detr/detr/models/setcriterion.py
@@ -247,7 +247,6 @@ def get_foreground_background_loss(
         return loss_map[loss](outputs, targets, indices, num_boxes, **kwargs)
 
     def _forward(self, outputs, outputs_without_aux, targets):
-
         # Retrieve the matching between the outputs of the last layer and the targets
         # A list where each item is [row_indices, col_indices]
         indices = self.matcher(outputs_without_aux, targets)
diff --git a/projects_oss/detr/detr/models/transformer.py b/projects_oss/detr/detr/models/transformer.py
index 7c21ffeb..3c6831f4 100644
--- a/projects_oss/detr/detr/models/transformer.py
+++ b/projects_oss/detr/detr/models/transformer.py
@@ -9,6 +9,7 @@
     * extra LN at the end of encoder is removed
     * decoder returns a stack of activations from all decoding layers
 """
+
 import copy
 from typing import Optional
 
diff --git a/projects_oss/detr/detr/util/box_ops.py b/projects_oss/detr/detr/util/box_ops.py
index e84e06e9..f9e22421 100644
--- a/projects_oss/detr/detr/util/box_ops.py
+++ b/projects_oss/detr/detr/util/box_ops.py
@@ -4,6 +4,7 @@
 """
 Utilities for bounding box manipulation and GIoU.
 """
+
 import torch
 from torchvision.ops.boxes import box_area
 
diff --git a/projects_oss/detr/detr/util/misc.py b/projects_oss/detr/detr/util/misc.py
index 0a1983ef..1f8c3af8 100644
--- a/projects_oss/detr/detr/util/misc.py
+++ b/projects_oss/detr/detr/util/misc.py
@@ -6,6 +6,7 @@
 
 Mostly copy-paste from torchvision references.
 """
+
 import datetime
 import os
 import pickle
diff --git a/projects_oss/detr/detr/util/plot_utils.py b/projects_oss/detr/detr/util/plot_utils.py
index 5bb06c69..4477d4df 100644
--- a/projects_oss/detr/detr/util/plot_utils.py
+++ b/projects_oss/detr/detr/util/plot_utils.py
@@ -3,6 +3,7 @@
 """
 Plotting utilities to visualize training logs.
 """
+
 from pathlib import Path, PurePath
 
 import matplotlib.pyplot as plt
diff --git a/projects_oss/detr/test_all.py b/projects_oss/detr/test_all.py
index b214d19a..08e90acd 100644
--- a/projects_oss/detr/test_all.py
+++ b/projects_oss/detr/test_all.py
@@ -195,7 +195,6 @@ def run_model(
             )
 
     def ort_validate(self, onnx_io, inputs, outputs, tolerate_small_mismatch=False):
-
         inputs, _ = torch.jit._flatten(inputs)
         outputs, _ = torch.jit._flatten(outputs)
 
diff --git a/projects_oss/detr/test_op.py b/projects_oss/detr/test_op.py
index 03695796..ca4345d8 100644
--- a/projects_oss/detr/test_op.py
+++ b/projects_oss/detr/test_op.py
@@ -120,7 +120,6 @@ def test_forward_equal_with_pytorch_float(self):
     def test_gradient_numerical(
         self, channels=4, grad_value=True, grad_sampling_loc=True, grad_attn_weight=True
     ):
-
         value = torch.rand(N, S, M, channels).cuda() * 0.01
         sampling_locations = torch.rand(N, Lq, M, L, P, 2).cuda()
         attention_weights = torch.rand(N, Lq, M, L, P).cuda() + 1e-5
diff --git a/tests/data/test_d2go_datasets.py b/tests/data/test_d2go_datasets.py
index a28caa13..4717d70d 100644
--- a/tests/data/test_d2go_datasets.py
+++ b/tests/data/test_d2go_datasets.py
@@ -303,7 +303,6 @@ def LineMetadata():
 
     @tempdir
     def test_coco_create_adhoc_class_to_use_dataset(self, tmp_dir):
-
         image_dir, json_file = create_test_images_and_dataset_json(
             tmp_dir, num_classes=2
         )
@@ -339,7 +338,6 @@ def test_coco_create_adhoc_class_to_use_dataset(self, tmp_dir):
 
     @tempdir
     def test_register_coco_dataset_registry(self, tmp_dir):
-
         dummy_buffer = []
 
         @COCO_REGISTER_FUNCTION_REGISTRY.register()
@@ -370,7 +368,6 @@ def _register_dummy_function_coco(dataset_name, split_dict):
 
     @tempdir
     def test_adhoc_register_coco_dataset_registry(self, tmp_dir):
-
         dummy_buffer = []
 
         def _dummy_load_func():
@@ -378,7 +375,6 @@ def _dummy_load_func():
 
         @COCO_REGISTER_FUNCTION_REGISTRY.register()
         def _register_dummy_function_coco_adhoc(dataset_name, split_dict):
-
             json_file = split_dict[ANN_FN]
             image_root = split_dict[IM_DIR]
 
diff --git a/tests/export/test_torchscript.py b/tests/export/test_torchscript.py
index ec44d27e..e9a5b0bb 100644
--- a/tests/export/test_torchscript.py
+++ b/tests/export/test_torchscript.py
@@ -14,7 +14,6 @@ def mock_export(cls, model, input_args, save_path, export_method, **export_kwarg
 
 class TestTorchscriptExportMethods(unittest.TestCase):
     def test_update_export_kwargs_from_export_method(self):
-
         _empty_export_kwargs = {}
 
         def try_mock_export(export_method: str, export_kwargs=_empty_export_kwargs):
diff --git a/tests/misc/test_config.py b/tests/misc/test_config.py
index a495956e..1e8eafac 100755
--- a/tests/misc/test_config.py
+++ b/tests/misc/test_config.py
@@ -258,7 +258,6 @@ def test_not_scale_for_zero_world_size(self):
 
 class TestConfigDefaultsGen(unittest.TestCase):
     def test_case1(self):
-
         # register in local scope
         @CONFIG_UPDATER_REGISTRY.register()
         def _test1(cfg):