fix flake8 issues (#554)

Summary: Pull Request resolved: #554

Reviewed By: NarineK

Differential Revision: D25418107

Pulled By: edward-io

fbshipit-source-id: 960ed22c5f6845ac9fedff8793196660d6fd5529

Author: edward-io

captum/insights/attr_vis/app.py

@@ -3,7 +3,6 @@
 from typing import (
     Any,
     Callable,
-    cast,
     Dict,
     Iterable,
     List,
@@ -150,8 +149,8 @@ def __init__(
         r"""
         Args:
 
-            models (torch.nn.module): One or more PyTorch modules (models) for attribution
-                          visualization.
+            models (torch.nn.module): One or more PyTorch modules (models) for
+                          attribution visualization.
             classes (list of string): List of strings corresponding to the names of
                           classes for classification.
             features (list of BaseFeature): List of BaseFeatures, which correspond
@@ -399,27 +398,25 @@ def _calculate_vis_output(
             # *an input contains multiple features that represent it
             #   e.g. all the pixels that describe an image is an input
 
-            attrs_per_input_feature = (
-                self.attribution_calculation.calculate_attribution(
-                    baselines,
-                    transformed_inputs,
-                    additional_forward_args,
-                    target,
-                    self._config.attribution_method,
-                    self._config.attribution_arguments,
-                    model,
-                )
+            attrs_per_feature = self.attribution_calculation.calculate_attribution(
+                baselines,
+                transformed_inputs,
+                additional_forward_args,
+                target,
+                self._config.attribution_method,
+                self._config.attribution_arguments,
+                model,
             )
 
             net_contrib = self.attribution_calculation.calculate_net_contrib(
-                attrs_per_input_feature
+                attrs_per_feature
             )
 
             # the features per input given
             features_per_input = [
                 feature.visualize(attr, data, contrib)
                 for feature, attr, data, contrib in zip(
-                    self.features, attrs_per_input_feature, inputs, net_contrib
+                    self.features, attrs_per_feature, inputs, net_contrib
                 )
             ]
 

captum/insights/attr_vis/attribution_calculation.py

@@ -49,11 +49,11 @@ def calculate_predicted_scores(
     ) -> Tuple[
         List[OutputScore], Optional[List[Tuple[Tensor, ...]]], Tuple[Tensor, ...]
     ]:
-        # Check to see if these inputs already have caches baselines and transformed inputs
-        hashableInputs = tuple(inputs)
-        if hashableInputs in self.baseline_cache:
-            baselines_group = self.baseline_cache[hashableInputs]
-            transformed_inputs = self.transformed_input_cache[hashableInputs]
+        # Check if inputs have cached baselines and transformed inputs
+        hashable_inputs = tuple(inputs)
+        if hashable_inputs in self.baseline_cache:
+            baselines_group = self.baseline_cache[hashable_inputs]
+            transformed_inputs = self.transformed_input_cache[hashable_inputs]
         else:
             # Initialize baselines
             baseline_transforms_len = 1  # todo support multiple baselines
@@ -79,8 +79,8 @@ def calculate_predicted_scores(
 
             baselines = cast(List[List[Optional[Tensor]]], baselines)
             baselines_group = [tuple(b) for b in baselines]
-            self.baseline_cache[hashableInputs] = baselines_group
-            self.transformed_input_cache[hashableInputs] = transformed_inputs
+            self.baseline_cache[hashable_inputs] = baselines_group
+            self.transformed_input_cache[hashable_inputs] = transformed_inputs
 
         outputs = _run_forward(
             model,