XX.

src/transformers/debug_utils.py

@@ -133,7 +133,9 @@ class DebugUnderflowOverflow:
 
     """
 
-    def __init__(self, model, max_frames_to_save=21, trace_batch_nums=[], abort_after_batch_num=None):
+    def __init__(self, model, max_frames_to_save=21, trace_batch_nums=None, abort_after_batch_num=None):
+        if trace_batch_nums is None:
+            trace_batch_nums = []
         self.model = model
         self.trace_batch_nums = trace_batch_nums
         self.abort_after_batch_num = abort_after_batch_num

src/transformers/utils/fx.py

@@ -235,7 +235,9 @@ class HFTracer(Tracer):
 
     default_methods_to_record = {"__bool__", "size", "dim"}
 
-    def __init__(self, batch_size=1, sequence_length=[128, 128], num_choices=-1):
+    def __init__(self, batch_size=1, sequence_length=None, num_choices=-1):
+        if sequence_length is None:
+            sequence_length = [128, 128]
         super().__init__()
 
         if not is_torch_fx_available():

tests/test_modeling_canine.py

@@ -363,7 +363,7 @@ class CanineModelTest(ModelTesterMixin, unittest.TestCase):
             t[t != t] = 0
             return t
 
-        def check_equivalence(model, tuple_inputs, dict_inputs, additional_kwargs={}):
+        def check_equivalence(model, tuple_inputs, dict_inputs, **additional_kwargs):
             with torch.no_grad():
                 tuple_output = model(**tuple_inputs, return_dict=False, **additional_kwargs)
                 dict_output = model(**dict_inputs, return_dict=True, **additional_kwargs).to_tuple()

tests/test_modeling_common.py

@@ -1366,7 +1366,7 @@ class ModelTesterMixin:
             t[t != t] = 0
             return t
 
-        def check_equivalence(model, tuple_inputs, dict_inputs, additional_kwargs={}):
+        def check_equivalence(model, tuple_inputs, dict_inputs, **additional_kwargs):
             with torch.no_grad():
                 tuple_output = model(**tuple_inputs, return_dict=False, **additional_kwargs)
                 dict_output = model(**dict_inputs, return_dict=True, **additional_kwargs).to_tuple()

tests/test_modeling_flax_common.py

@@ -137,7 +137,7 @@ class FlaxModelTesterMixin:
             t[t != t] = 0
             return t
 
-        def check_equivalence(model, tuple_inputs, dict_inputs, additional_kwargs={}):
+        def check_equivalence(model, tuple_inputs, dict_inputs, **additional_kwargs):
             tuple_output = model(**tuple_inputs, return_dict=False, **additional_kwargs)
             dict_output = model(**dict_inputs, return_dict=True, **additional_kwargs).to_tuple()
 

tests/test_modeling_fnet.py

@@ -313,7 +313,7 @@ class FNetModelTest(ModelTesterMixin, unittest.TestCase):
             t[t != t] = 0
             return t
 
-        def check_equivalence(model, tuple_inputs, dict_inputs, additional_kwargs={}):
+        def check_equivalence(model, tuple_inputs, dict_inputs, **additional_kwargs):
             with torch.no_grad():
                 tuple_output = model(**tuple_inputs, return_dict=False, **additional_kwargs)
                 dict_output = model(**dict_inputs, return_dict=True, **additional_kwargs).to_tuple()

tests/test_modeling_tf_common.py

@@ -875,7 +875,7 @@ class TFModelTesterMixin:
 
         config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
 
-        def check_equivalence(model, tuple_inputs, dict_inputs, additional_kwargs={}):
+        def check_equivalence(model, tuple_inputs, dict_inputs, **additional_kwargs):
             tuple_output = model(tuple_inputs, return_dict=False, **additional_kwargs)
             dict_output = model(dict_inputs, return_dict=True, **additional_kwargs).to_tuple()