Address pylint warnings

Author: pwolnows

tools/accuracy_checker/accuracy_checker/annotation_converters/sentence_similarity.py

@@ -150,7 +150,7 @@ def convert_single_example(self, example, min_score=None, max_score=None):
                 tokens.append("[CLS]" if self.support_vocab else CLS_ID)
                 segment_ids.append(SEG_ID_CLS)
         else:
-            tokens = self.tokenizer.tokenize(example.text, add_special_tokens=True)
+            tokens = self.tokenizer.tokenize(example.text)
             segment_ids = [SEG_ID_A] * len(tokens)
 
             if len(tokens) > self.max_seq_length:

tools/accuracy_checker/accuracy_checker/annotation_converters/text_classification.py

@@ -156,8 +156,8 @@ def convert_single_example(self, example): # pylint:disable=R0912
                 segment_ids.append(SEG_ID_CLS)
         else:
             if example.text_b:
-                tokens = self.tokenizer.tokenize((example.text_a, example.text_b), add_special_tokens=True)
-                len_tokens_a = len(self.tokenizer.tokenize(example.text_a, add_special_tokens=True))
+                tokens = self.tokenizer.tokenize((example.text_a, example.text_b))
+                len_tokens_a = len(self.tokenizer.tokenize(example.text_a))
                 segment_ids = [SEG_ID_A] * len_tokens_a + [SEG_ID_B] * (len(tokens) - len_tokens_a)
             else:
                 tokens = self.tokenizer.tokenize(example.text_a)

tools/accuracy_checker/accuracy_checker/launcher/launcher.py

@@ -263,8 +263,7 @@ def _provide_inputs_info_to_meta(self, meta):
 
         return meta
 
-    @staticmethod
-    def fit_to_input(data, layer_name, layout, precision, template=None):
+    def fit_to_input(self, data, layer_name, layout, precision, template=None):
         layout_used = False
         if layout is not None and len(np.shape(data)) == len(layout):
             data = np.transpose(data, layout)

tools/accuracy_checker/accuracy_checker/launcher/pytorch_launcher.py

@@ -156,7 +156,7 @@ def load_module(self, model_cls, module_args, module_kwargs, checkpoint=None, st
         model_cls = module_parts[-1]
         model_path = ".".join(module_parts[:-1])
         with append_to_path(python_path):
-            model_cls = importlib.import_module(model_path).__getattribute__(model_cls)
+            model_cls = getattr(importlib.import_module(model_path), model_cls)
             module = model_cls(*module_args, **module_kwargs)
             if init_method is not None:
                 if hasattr(model_cls, init_method):
@@ -247,16 +247,15 @@ def _convert_to_numpy(self, input_dict):
     def forward(self, outputs):
         if hasattr(outputs, 'logits') and 'logits' in self.output_names:
             return {'logits': outputs.logits}
-        elif hasattr(outputs, 'last_hidden_state') and 'last_hidden_state' in self.output_names:
+        if hasattr(outputs, 'last_hidden_state') and 'last_hidden_state' in self.output_names:
             return {'last_hidden_state': outputs.last_hidden_state}
         return list(outputs)
 
     def predict(self, inputs, metadata=None, **kwargs):
         results = []
         with self._torch.no_grad():
             for batch_input in inputs:
-                if metadata[0].get('input_is_dict_type') or \
-                    (isinstance(batch_input, dict) and 'input' in batch_input):
+                if metadata[0].get('input_is_dict_type') or (isinstance(batch_input, dict) and 'input' in batch_input):
                     outputs = self.module(batch_input['input'])
                 else:
                     outputs = self.module(**batch_input)