Sign upLogin

tensorflow/models

View on GitHub
Star
official/nlp/modeling/layers/reuse_transformer.py

Summary

Maintainability
F
4 days
Test Coverage

Function call has a Cognitive Complexity of 29 (exceeds 5 allowed). Consider refactoring.
Open

  def call(self, inputs):
    """Transformer self-attention encoder block call.

    Args:
      inputs: a single tensor or a list of tensors.
Severity: Minor
Found in official/nlp/modeling/layers/reuse_transformer.py - About 4 hrs to fix

Cognitive Complexity

Cognitive Complexity is a measure of how difficult a unit of code is to intuitively understand. Unlike Cyclomatic Complexity, which determines how difficult your code will be to test, Cognitive Complexity tells you how difficult your code will be to read and comprehend.

A method's cognitive complexity is based on a few simple rules:

  • Code is not considered more complex when it uses shorthand that the language provides for collapsing multiple statements into one
  • Code is considered more complex for each "break in the linear flow of the code"
  • Code is considered more complex when "flow breaking structures are nested"

Further reading

File reuse_transformer.py has 318 lines of code (exceeds 250 allowed). Consider refactoring.
Open

# Copyright 2024 The TensorFlow Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
Severity: Minor
Found in official/nlp/modeling/layers/reuse_transformer.py - About 3 hrs to fix

    Function __init__ has 26 arguments (exceeds 4 allowed). Consider refactoring.
    Open

      def __init__(self,
    Severity: Major
    Found in official/nlp/modeling/layers/reuse_transformer.py - About 3 hrs to fix

      Function __init__ has 30 lines of code (exceeds 25 allowed). Consider refactoring.
      Open

        def __init__(self,
               num_attention_heads,
               inner_dim,
               inner_activation,
               head_size=None,
      Severity: Minor
      Found in official/nlp/modeling/layers/reuse_transformer.py - About 1 hr to fix

        Function build has a Cognitive Complexity of 9 (exceeds 5 allowed). Consider refactoring.
        Open

          def build(self, input_shape):
    if isinstance(input_shape, tf.TensorShape):
      input_tensor_shape = input_shape
    elif isinstance(input_shape, (list, tuple)):
      input_tensor_shape = tf.TensorShape(input_shape[0])
        Severity: Minor
        Found in official/nlp/modeling/layers/reuse_transformer.py - About 55 mins to fix

        Cognitive Complexity

        Cognitive Complexity is a measure of how difficult a unit of code is to intuitively understand. Unlike Cyclomatic Complexity, which determines how difficult your code will be to test, Cognitive Complexity tells you how difficult your code will be to read and comprehend.

        A method's cognitive complexity is based on a few simple rules:

        • Code is not considered more complex when it uses shorthand that the language provides for collapsing multiple statements into one
        • Code is considered more complex for each "break in the linear flow of the code"
        • Code is considered more complex when "flow breaking structures are nested"

        Further reading

        Function __init__ has a Cognitive Complexity of 6 (exceeds 5 allowed). Consider refactoring.
        Open

          def __init__(self,
               num_attention_heads,
               inner_dim,
               inner_activation,
               head_size=None,
        Severity: Minor
        Found in official/nlp/modeling/layers/reuse_transformer.py - About 25 mins to fix

        Cognitive Complexity

        Cognitive Complexity is a measure of how difficult a unit of code is to intuitively understand. Unlike Cyclomatic Complexity, which determines how difficult your code will be to test, Cognitive Complexity tells you how difficult your code will be to read and comprehend.

        A method's cognitive complexity is based on a few simple rules:

        • Code is not considered more complex when it uses shorthand that the language provides for collapsing multiple statements into one
        • Code is considered more complex for each "break in the linear flow of the code"
        • Code is considered more complex when "flow breaking structures are nested"

        Further reading

        Similar blocks of code found in 2 locations. Consider refactoring.
        Open

            if isinstance(inputs, (list, tuple)):
      if len(inputs) == 2:
        input_tensor, attention_mask = inputs
        reuse_attention_scores = None
      elif len(inputs) == 3:
        Severity: Major
        Found in official/nlp/modeling/layers/reuse_transformer.py and 1 other location - About 6 hrs to fix
        official/nlp/modeling/layers/transformer_scaffold.py on lines 295..305

        Duplicated Code

        Duplicated code can lead to software that is hard to understand and difficult to change. The Don't Repeat Yourself (DRY) principle states:

        Every piece of knowledge must have a single, unambiguous, authoritative representation within a system.

        When you violate DRY, bugs and maintenance problems are sure to follow. Duplicated code has a tendency to both continue to replicate and also to diverge (leaving bugs as two similar implementations differ in subtle ways).

        Tuning

        This issue has a mass of 100.

        We set useful threshold defaults for the languages we support but you may want to adjust these settings based on your project guidelines.

        The threshold configuration represents the minimum mass a code block must have to be analyzed for duplication. The lower the threshold, the more fine-grained the comparison.

        If the engine is too easily reporting duplication, try raising the threshold. If you suspect that the engine isn't catching enough duplication, try lowering the threshold. The best setting tends to differ from language to language.

        See codeclimate-duplication's documentation for more information about tuning the mass threshold in your .codeclimate.yml.

        Refactorings

        Further Reading

        Identical blocks of code found in 3 locations. Consider refactoring.
        Open

            if isinstance(input_shape, tf.TensorShape):
      input_tensor_shape = input_shape
    elif isinstance(input_shape, (list, tuple)):
      input_tensor_shape = tf.TensorShape(input_shape[0])
    else:
        Severity: Major
        Found in official/nlp/modeling/layers/reuse_transformer.py and 2 other locations - About 3 hrs to fix
        official/nlp/modeling/layers/transformer_scaffold.py on lines 128..135
        official/projects/detr/modeling/transformer.py on lines 244..251

        Duplicated Code

        Duplicated code can lead to software that is hard to understand and difficult to change. The Don't Repeat Yourself (DRY) principle states:

        Every piece of knowledge must have a single, unambiguous, authoritative representation within a system.

        When you violate DRY, bugs and maintenance problems are sure to follow. Duplicated code has a tendency to both continue to replicate and also to diverge (leaving bugs as two similar implementations differ in subtle ways).

        Tuning

        This issue has a mass of 64.

        We set useful threshold defaults for the languages we support but you may want to adjust these settings based on your project guidelines.

        The threshold configuration represents the minimum mass a code block must have to be analyzed for duplication. The lower the threshold, the more fine-grained the comparison.

        If the engine is too easily reporting duplication, try raising the threshold. If you suspect that the engine isn't catching enough duplication, try lowering the threshold. The best setting tends to differ from language to language.

        See codeclimate-duplication's documentation for more information about tuning the mass threshold in your .codeclimate.yml.

        Refactorings

        Further Reading

        Identical blocks of code found in 2 locations. Consider refactoring.
        Open

              if self._norm_first:
        source_tensor = input_tensor[:, 0:self._output_range, :]
        input_tensor = self._attention_layer_norm(input_tensor)
        if key_value is not None:
          key_value = self._attention_layer_norm(key_value)
        Severity: Major
        Found in official/nlp/modeling/layers/reuse_transformer.py and 1 other location - About 2 hrs to fix
        official/projects/detr/modeling/transformer.py on lines 378..382

        Duplicated Code

        Duplicated code can lead to software that is hard to understand and difficult to change. The Don't Repeat Yourself (DRY) principle states:

        Every piece of knowledge must have a single, unambiguous, authoritative representation within a system.

        When you violate DRY, bugs and maintenance problems are sure to follow. Duplicated code has a tendency to both continue to replicate and also to diverge (leaving bugs as two similar implementations differ in subtle ways).

        Tuning

        This issue has a mass of 57.

        We set useful threshold defaults for the languages we support but you may want to adjust these settings based on your project guidelines.

        The threshold configuration represents the minimum mass a code block must have to be analyzed for duplication. The lower the threshold, the more fine-grained the comparison.

        If the engine is too easily reporting duplication, try raising the threshold. If you suspect that the engine isn't catching enough duplication, try lowering the threshold. The best setting tends to differ from language to language.

        See codeclimate-duplication's documentation for more information about tuning the mass threshold in your .codeclimate.yml.

        Refactorings

        Further Reading

        Identical blocks of code found in 2 locations. Consider refactoring.
        Open

            if self._output_range:
      if self._norm_first:
        source_tensor = input_tensor[:, 0:self._output_range, :]
        input_tensor = self._attention_layer_norm(input_tensor)
        if key_value is not None:
        Severity: Major
        Found in official/nlp/modeling/layers/reuse_transformer.py and 1 other location - About 1 hr to fix
        official/projects/detr/modeling/transformer.py on lines 377..392

        Duplicated Code

        Duplicated code can lead to software that is hard to understand and difficult to change. The Don't Repeat Yourself (DRY) principle states:

        Every piece of knowledge must have a single, unambiguous, authoritative representation within a system.

        When you violate DRY, bugs and maintenance problems are sure to follow. Duplicated code has a tendency to both continue to replicate and also to diverge (leaving bugs as two similar implementations differ in subtle ways).

        Tuning

        This issue has a mass of 49.

        We set useful threshold defaults for the languages we support but you may want to adjust these settings based on your project guidelines.

        The threshold configuration represents the minimum mass a code block must have to be analyzed for duplication. The lower the threshold, the more fine-grained the comparison.

        If the engine is too easily reporting duplication, try raising the threshold. If you suspect that the engine isn't catching enough duplication, try lowering the threshold. The best setting tends to differ from language to language.

        See codeclimate-duplication's documentation for more information about tuning the mass threshold in your .codeclimate.yml.

        Refactorings

        Further Reading

        Similar blocks of code found in 2 locations. Consider refactoring.
        Open

            self._intermediate_dense = tf_keras.layers.EinsumDense(
        einsum_equation,
        output_shape=(None, self._inner_dim),
        bias_axes="d",
        kernel_initializer=tf_utils.clone_initializer(self._kernel_initializer),
        Severity: Minor
        Found in official/nlp/modeling/layers/reuse_transformer.py and 1 other location - About 55 mins to fix
        official/nlp/modeling/layers/transformer.py on lines 321..326

        Duplicated Code

        Duplicated code can lead to software that is hard to understand and difficult to change. The Don't Repeat Yourself (DRY) principle states:

        Every piece of knowledge must have a single, unambiguous, authoritative representation within a system.

        When you violate DRY, bugs and maintenance problems are sure to follow. Duplicated code has a tendency to both continue to replicate and also to diverge (leaving bugs as two similar implementations differ in subtle ways).

        Tuning

        This issue has a mass of 37.

        We set useful threshold defaults for the languages we support but you may want to adjust these settings based on your project guidelines.

        The threshold configuration represents the minimum mass a code block must have to be analyzed for duplication. The lower the threshold, the more fine-grained the comparison.

        If the engine is too easily reporting duplication, try raising the threshold. If you suspect that the engine isn't catching enough duplication, try lowering the threshold. The best setting tends to differ from language to language.

        See codeclimate-duplication's documentation for more information about tuning the mass threshold in your .codeclimate.yml.

        Refactorings

        Further Reading

        Similar blocks of code found in 7 locations. Consider refactoring.
        Open

              if hidden_size % self._num_heads != 0:
        raise ValueError(
            "The input size (%d) is not a multiple of the number of attention "
            "heads (%d)" % (hidden_size, self._num_heads))
        Severity: Major
        Found in official/nlp/modeling/layers/reuse_transformer.py and 6 other locations - About 50 mins to fix
        official/nlp/modeling/layers/tn_transformer_expand_condense.py on lines 127..130
        official/nlp/modeling/layers/transformer.py on lines 260..263
        official/nlp/modeling/layers/transformer_scaffold.py on lines 142..145
        official/nlp/modeling/layers/transformer_xl.py on lines 142..145
        official/projects/detr/modeling/transformer.py on lines 256..259
        official/projects/detr/modeling/transformer.py on lines 670..673

        Duplicated Code

        Duplicated code can lead to software that is hard to understand and difficult to change. The Don't Repeat Yourself (DRY) principle states:

        Every piece of knowledge must have a single, unambiguous, authoritative representation within a system.

        When you violate DRY, bugs and maintenance problems are sure to follow. Duplicated code has a tendency to both continue to replicate and also to diverge (leaving bugs as two similar implementations differ in subtle ways).

        Tuning

        This issue has a mass of 36.

        We set useful threshold defaults for the languages we support but you may want to adjust these settings based on your project guidelines.

        The threshold configuration represents the minimum mass a code block must have to be analyzed for duplication. The lower the threshold, the more fine-grained the comparison.

        If the engine is too easily reporting duplication, try raising the threshold. If you suspect that the engine isn't catching enough duplication, try lowering the threshold. The best setting tends to differ from language to language.

        See codeclimate-duplication's documentation for more information about tuning the mass threshold in your .codeclimate.yml.

        Refactorings

        Further Reading

        Identical blocks of code found in 5 locations. Consider refactoring.
        Open

            if attention_initializer:
      self._attention_initializer = tf_keras.initializers.get(
          attention_initializer)
    else:
      self._attention_initializer = tf_utils.clone_initializer(
        Severity: Major
        Found in official/nlp/modeling/layers/reuse_transformer.py and 4 other locations - About 45 mins to fix
        official/nlp/modeling/layers/tn_transformer_expand_condense.py on lines 100..105
        official/nlp/modeling/layers/transformer.py on lines 234..239
        official/projects/detr/modeling/transformer.py on lines 235..240
        official/projects/detr/modeling/transformer.py on lines 656..661

        Duplicated Code

        Duplicated code can lead to software that is hard to understand and difficult to change. The Don't Repeat Yourself (DRY) principle states:

        Every piece of knowledge must have a single, unambiguous, authoritative representation within a system.

        When you violate DRY, bugs and maintenance problems are sure to follow. Duplicated code has a tendency to both continue to replicate and also to diverge (leaving bugs as two similar implementations differ in subtle ways).

        Tuning

        This issue has a mass of 35.

        We set useful threshold defaults for the languages we support but you may want to adjust these settings based on your project guidelines.

        The threshold configuration represents the minimum mass a code block must have to be analyzed for duplication. The lower the threshold, the more fine-grained the comparison.

        If the engine is too easily reporting duplication, try raising the threshold. If you suspect that the engine isn't catching enough duplication, try lowering the threshold. The best setting tends to differ from language to language.

        See codeclimate-duplication's documentation for more information about tuning the mass threshold in your .codeclimate.yml.

        Refactorings

        Further Reading

        Similar blocks of code found in 10 locations. Consider refactoring.
        Open

            if self._norm_first:
      attention_output = source_tensor + attention_output
    else:
      attention_output = self._attention_layer_norm(target_tensor +
        Severity: Major
        Found in official/nlp/modeling/layers/reuse_transformer.py and 9 other locations - About 45 mins to fix
        official/nlp/modeling/layers/tn_transformer_expand_condense.py on lines 235..238
        official/nlp/modeling/layers/tn_transformer_expand_condense.py on lines 250..253
        official/nlp/modeling/layers/transformer.py on lines 419..423
        official/nlp/modeling/layers/transformer.py on lines 438..442
        official/nlp/modeling/layers/transformer.py on lines 453..456
        official/projects/detr/modeling/transformer.py on lines 402..405
        official/projects/detr/modeling/transformer.py on lines 814..818
        official/projects/detr/modeling/transformer.py on lines 830..834
        official/projects/detr/modeling/transformer.py on lines 845..848

        Duplicated Code

        Duplicated code can lead to software that is hard to understand and difficult to change. The Don't Repeat Yourself (DRY) principle states:

        Every piece of knowledge must have a single, unambiguous, authoritative representation within a system.

        When you violate DRY, bugs and maintenance problems are sure to follow. Duplicated code has a tendency to both continue to replicate and also to diverge (leaving bugs as two similar implementations differ in subtle ways).

        Tuning

        This issue has a mass of 35.

        We set useful threshold defaults for the languages we support but you may want to adjust these settings based on your project guidelines.

        The threshold configuration represents the minimum mass a code block must have to be analyzed for duplication. The lower the threshold, the more fine-grained the comparison.

        If the engine is too easily reporting duplication, try raising the threshold. If you suspect that the engine isn't catching enough duplication, try lowering the threshold. The best setting tends to differ from language to language.

        See codeclimate-duplication's documentation for more information about tuning the mass threshold in your .codeclimate.yml.

        Refactorings

        Further Reading

        Identical blocks of code found in 2 locations. Consider refactoring.
        Open

            self._output_dense = tf_keras.layers.EinsumDense(
        einsum_equation,
        output_shape=(None, hidden_size),
        bias_axes="d",
        name="output",
        Severity: Minor
        Found in official/nlp/modeling/layers/reuse_transformer.py and 1 other location - About 45 mins to fix
        official/nlp/modeling/layers/transformer_scaffold.py on lines 237..244

        Duplicated Code

        Duplicated code can lead to software that is hard to understand and difficult to change. The Don't Repeat Yourself (DRY) principle states:

        Every piece of knowledge must have a single, unambiguous, authoritative representation within a system.

        When you violate DRY, bugs and maintenance problems are sure to follow. Duplicated code has a tendency to both continue to replicate and also to diverge (leaving bugs as two similar implementations differ in subtle ways).

        Tuning

        This issue has a mass of 35.

        We set useful threshold defaults for the languages we support but you may want to adjust these settings based on your project guidelines.

        The threshold configuration represents the minimum mass a code block must have to be analyzed for duplication. The lower the threshold, the more fine-grained the comparison.

        If the engine is too easily reporting duplication, try raising the threshold. If you suspect that the engine isn't catching enough duplication, try lowering the threshold. The best setting tends to differ from language to language.

        See codeclimate-duplication's documentation for more information about tuning the mass threshold in your .codeclimate.yml.

        Refactorings

        Further Reading

        Identical blocks of code found in 7 locations. Consider refactoring.
        Open

            common_kwargs = dict(
        kernel_regularizer=self._kernel_regularizer,
        bias_regularizer=self._bias_regularizer,
        activity_regularizer=self._activity_regularizer,
        kernel_constraint=self._kernel_constraint,
        Severity: Major
        Found in official/nlp/modeling/layers/reuse_transformer.py and 6 other locations - About 35 mins to fix
        official/nlp/modeling/layers/gated_feedforward.py on lines 103..108
        official/nlp/modeling/layers/multi_channel_attention.py on lines 63..68
        official/nlp/modeling/layers/reuse_attention.py on lines 351..356
        official/nlp/modeling/layers/tn_transformer_expand_condense.py on lines 132..137
        official/nlp/modeling/layers/transformer.py on lines 265..270
        official/nlp/modeling/layers/transformer_scaffold.py on lines 148..153

        Duplicated Code

        Duplicated code can lead to software that is hard to understand and difficult to change. The Don't Repeat Yourself (DRY) principle states:

        Every piece of knowledge must have a single, unambiguous, authoritative representation within a system.

        When you violate DRY, bugs and maintenance problems are sure to follow. Duplicated code has a tendency to both continue to replicate and also to diverge (leaving bugs as two similar implementations differ in subtle ways).

        Tuning

        This issue has a mass of 33.

        We set useful threshold defaults for the languages we support but you may want to adjust these settings based on your project guidelines.

        The threshold configuration represents the minimum mass a code block must have to be analyzed for duplication. The lower the threshold, the more fine-grained the comparison.

        If the engine is too easily reporting duplication, try raising the threshold. If you suspect that the engine isn't catching enough duplication, try lowering the threshold. The best setting tends to differ from language to language.

        See codeclimate-duplication's documentation for more information about tuning the mass threshold in your .codeclimate.yml.

        Refactorings

        Further Reading

        There are no issues that match your filters.

        Category
        Status