Black

pytorch_toolbelt/inference/ensembling.py

@@ -92,13 +92,13 @@ def __init__(self, models: List[nn.Module], reduction: str = "mean", outputs: Op
     def forward(self, *input, **kwargs):  # skipcq: PYL-W0221
         outputs = [model(*input, **kwargs) for model in self.models]
         output_is_dict = isinstance(outputs[0], dict)
-        output_is_list = isinstance(outputs[0], (list, tuple))
+        output_is_list = isinstance(outputs[0], (list, tuple)) # noqa
 
         if self.return_some_outputs:
             keys = self.outputs
-        elif isinstance(outputs[0], dict):
+        elif output_is_dict:
             keys = outputs[0].keys()
-        elif isinstance(outputs[0], (list, tuple)):
+        elif output_is_list:
             keys = list(range(len(outputs[0])))
         elif torch.is_tensor(outputs[0]):
             keys = None

pytorch_toolbelt/inference/tiles.py

@@ -1,6 +1,7 @@
 """Implementation of tile-based inference allowing to predict huge images that does not fit into GPU memory entirely
 in a sliding-window fashion and merging prediction mask back to full-resolution.
 """
+
 import dataclasses
 import math
 from typing import List, Iterable, Tuple, Union, Sequence

pytorch_toolbelt/inference/tta.py

@@ -3,6 +3,7 @@
 Despite this is called test-time augmentation, these method can be used at training time as well since all
 transformation written in PyTorch and respect gradients flow.
 """
+
 from collections import defaultdict
 from functools import partial
 from typing import Tuple, List, Optional, Union, Callable, Dict, Mapping

pytorch_toolbelt/losses/functional.py

@@ -29,7 +29,7 @@ def focal_loss_with_logits(
     ignore_index=None,
     activation: str = "sigmoid",
     softmax_dim: Optional[int] = None,
-    class_weights: Optional[torch.Tensor] = None
+    class_weights: Optional[torch.Tensor] = None,
 ) -> torch.Tensor:
     """Compute binary focal loss between target and output logits.
 
@@ -70,7 +70,9 @@ def focal_loss_with_logits(
     if reduced_threshold is None:
         focal_term = (1.0 - pt).pow(gamma)
     else:
-        focal_term = ((1.0 - pt) / (1 - reduced_threshold)).pow(gamma) #the focal term continuity breaks when reduced_threshold not equal to 0.5. At pt equal to reduced_threshold, the value of piecewise function of focal term should be 1 from both sides .
+        focal_term = ((1.0 - pt) / (1 - reduced_threshold)).pow(
+            gamma
+        )  # the focal term continuity breaks when reduced_threshold not equal to 0.5. At pt equal to reduced_threshold, the value of piecewise function of focal term should be 1 from both sides .
         focal_term = torch.masked_fill(focal_term, pt < reduced_threshold, 1)
 
     loss = focal_term * ce_loss

pytorch_toolbelt/modules/__init__.py

@@ -18,4 +18,4 @@
 from .initialization import *
 from .normalization import *
 
-from .heads import *
+from .heads import *

pytorch_toolbelt/modules/activations.py

@@ -255,7 +255,7 @@ def get_activation_block(activation_name: str):
         ACT_SWISH: Swish,
         ACT_SWISH_NAIVE: SwishNaive,
         ACT_SIGMOID: nn.Sigmoid,
-        ACT_SOFTMAX: nn.Softmax
+        ACT_SOFTMAX: nn.Softmax,
     }
 
     return ACTIVATIONS[activation_name.lower()]

pytorch_toolbelt/modules/backbone/senet.py

@@ -2,6 +2,7 @@
 ResNet code gently borrowed from
 https://github.com/pytorch/vision/blob/master/torchvision/models/resnet.py
 """
+
 from __future__ import print_function, division, absolute_import
 
 import math

pytorch_toolbelt/modules/encoders/__init__.py

@@ -2,6 +2,7 @@
 
 Encodes listed here provides easy way to swap backbone of classification/segmentation/detection model.
 """
+
 from .common import *
 from .densenet import *
 from .hrnet import *

pytorch_toolbelt/modules/encoders/common.py

@@ -2,6 +2,7 @@
 
 Encodes listed here provides easy way to swap backbone of classification/segmentation/detection model.
 """
+
 import math
 import warnings
 from typing import List, Union, Tuple, Iterable, Any

pytorch_toolbelt/modules/encoders/seresnet.py

@@ -2,6 +2,7 @@
 
 Encodes listed here provides easy way to swap backbone of classification/segmentation/detection model.
 """
+
 from typing import List
 
 import torch

pytorch_toolbelt/modules/encoders/timm/maxvit.py

@@ -1,11 +1,12 @@
 from .common import GenericTimmEncoder
 
+
 class MaxVitEncoder(GenericTimmEncoder):
-    def __init__(self, model_name:str, pretrained=True, **kwargs):
+    def __init__(self, model_name: str, pretrained=True, **kwargs):
         super().__init__(model_name, pretrained=pretrained, **kwargs)
 
     def change_input_channels(self, input_channels: int, mode="auto", **kwargs):
         from pytorch_toolbelt.modules import make_n_channel_input
 
         self.encoder.stem.conv1 = make_n_channel_input(self.encoder.stem.conv1, input_channels)
-        return self
+        return self

pytorch_toolbelt/modules/encoders/timm/resnet.py

@@ -181,6 +181,7 @@ def change_input_channels(self, input_channels: int, mode="auto", **kwargs):
         self.encoder.conv1[0] = make_n_channel_input(self.encoder.conv1[0], input_channels, mode=mode, **kwargs)
         return self
 
+
 class TimmResnet50D(GenericTimmEncoder):
     def __init__(
         self, pretrained=True, layers=None, activation=ACT_RELU, first_conv_stride_one: bool = False, **kwargs

pytorch_toolbelt/modules/heads/__init__.py

@@ -3,4 +3,4 @@
 from .deep_supervision import *
 from .hypercolumn import *
 from .segformer_head import *
-from .classification_heads import *
+from .classification_heads import *

pytorch_toolbelt/modules/interfaces.py

@@ -61,8 +61,7 @@ class HasInputFeaturesSpecification(Protocol):
     """
 
     @torch.jit.unused
-    def get_input_spec(self) -> FeatureMapsSpecification:
-        ...
+    def get_input_spec(self) -> FeatureMapsSpecification: ...
 
 
 class HasOutputFeaturesSpecification(Protocol):
@@ -71,8 +70,7 @@ class HasOutputFeaturesSpecification(Protocol):
     """
 
     @torch.jit.unused
-    def get_output_spec(self) -> FeatureMapsSpecification:
-        ...
+    def get_output_spec(self) -> FeatureMapsSpecification: ...
 
 
 class AbstractEncoder(nn.Module, HasOutputFeaturesSpecification):
@@ -108,8 +106,7 @@ def __init__(self, input_spec: FeatureMapsSpecification):
     @abstractmethod
     def forward(
         self, feature_maps: List[Tensor], output_size: Union[Tuple[int, int], torch.Size, None] = None
-    ) -> Union[Tensor, Tuple[Tensor, ...], List[Tensor], Mapping[str, Tensor]]:
-        ...
+    ) -> Union[Tensor, Tuple[Tensor, ...], List[Tensor], Mapping[str, Tensor]]: ...
 
     @torch.jit.unused
     def apply_to_final_layer(self, func: Callable[[nn.Module], None]):

pytorch_toolbelt/modules/pooling.py

@@ -200,16 +200,7 @@ def forward(self, x: Tensor) -> Tensor:
 
     def __repr__(self):
         p = torch.softplus(self.p) + 1
-        return (
-            self.__class__.__name__
-            + "("
-            + "p="
-            + "{:.4f}".format(p.item())
-            + ", "
-            + "eps="
-            + str(self.eps)
-            + ")"
-        )
+        return self.__class__.__name__ + "(" + "p=" + "{:.4f}".format(p.item()) + ", " + "eps=" + str(self.eps) + ")"
 
 
 class GlobalMaxAvgPooling2d(nn.Module):

pytorch_toolbelt/utils/random_utils.py

@@ -1,6 +1,7 @@
 """Utility functions to make your experiments reproducible
 
 """
+
 import random
 import warnings
 

pytorch_toolbelt/utils/torch_utils.py

@@ -198,7 +198,7 @@ def container_to_tensor(value: Union[np.ndarray, List, Tuple, Mapping, Any]):
         cls = type(value)
         return cls((k, container_to_tensor(v)) for k, v in value.items())
 
-    raise ValueError(f"Unsupported container type")
+    raise ValueError(f"Unsupported container type {type(value)}")
 
 
 def image_to_tensor(image: np.ndarray, dummy_channels_dim=True) -> torch.Tensor:

setup.py

@@ -169,7 +169,7 @@ def get_test_requirements():
         "Topic :: Scientific/Engineering :: Artificial Intelligence",
         "Topic :: Software Development :: Libraries",
         "Topic :: Software Development :: Libraries :: Python Modules",
-        "Topic :: Software Development :: Libraries :: Application Frameworks"
+        "Topic :: Software Development :: Libraries :: Application Frameworks",
         # "Private :: Do Not Upload"
     ],
 )