Sign upLogin

IBM/pytorchpipe

View on GitHub
Star
ptp/workers/online_trainer.py

Summary

Maintainability
F
5 days
Test Coverage

Function run_experiment has a Cognitive Complexity of 73 (exceeds 5 allowed). Consider refactoring.
Open

    def run_experiment(self):
        """
        Main function of the ``OnlineTrainer``, runs the experiment.

        Iterates over the (cycled) DataLoader (one iteration = one episode).
Severity: Minor
Found in ptp/workers/online_trainer.py - About 1 day 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

Avoid deeply nested control flow statements.
Open

                            for name, param in self.pipeline.named_parameters():
                                try:
                                    self.training_batch_writer.add_histogram(name + '/grad', 
                                        param.grad.data.cpu().numpy(), self.app_state.episode, bins='doane')
Severity: Major
Found in ptp/workers/online_trainer.py - About 45 mins to fix

    Function setup_experiment has a Cognitive Complexity of 8 (exceeds 5 allowed). Consider refactoring.
    Open

        def setup_experiment(self):
        """
        Sets up experiment for episode trainer:

            - Calls base class setup_experiment to parse the command line arguments,
    Severity: Minor
    Found in ptp/workers/online_trainer.py - About 45 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

    Avoid deeply nested control flow statements.
    Open

                                if (validation_batch_loss < self.loss_stop_threshold):
                                # Change the status.
                                training_status = "Converged (Partial Validation Loss went below " \
                                    "Loss Stop threshold {})".format(self.loss_stop_threshold)
    Severity: Major
    Found in ptp/workers/online_trainer.py - About 45 mins to fix

      Avoid deeply nested control flow statements.
      Open

                                  for name, param in self.pipeline.named_parameters():
                                try:
                                    self.training_batch_writer.add_histogram(name, 
                                        param.data.cpu().numpy(), self.app_state.episode, bins='doane')
      Severity: Major
      Found in ptp/workers/online_trainer.py - About 45 mins to fix

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

                            if (self.training_batch_writer is not None) and \
                            (self.app_state.episode % self.app_state.args.logging_interval == 0):
                        self.training_stat_col.export_to_tensorboard()

                        # Export histograms.
        Severity: Major
        Found in ptp/workers/online_trainer.py and 1 other location - About 2 days to fix
        ptp/workers/offline_trainer.py on lines 218..240

        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 234.

        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

                                if self.curric_done or not self.must_finish_curriculum:

                            # Check the Partial Validation loss.
                            if (validation_batch_loss < self.loss_stop_threshold):
                                # Change the status.
        Severity: Major
        Found in ptp/workers/online_trainer.py and 1 other location - About 3 hrs to fix
        ptp/workers/offline_trainer.py on lines 290..301

        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

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

                if self.episode_limit <= 0:
            self.logger.error("OnLine Trainer relies on episodes, thus 'episode_limit' must be a positive number!")
            exit(-5)
        else:
            log_str += "   IV: Setting the Episode Limit to: {}\n".format(self.episode_limit)
        Severity: Major
        Found in ptp/workers/online_trainer.py and 1 other location - About 1 hr to fix
        ptp/workers/offline_trainer.py on lines 83..87

        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 46.

        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 3 locations. Consider refactoring.
        Open

                if self.epoch_limit <= 0:
            log_str += "  III: Termination based on Epoch Limit is disabled\n"
            # Set to infinity.
            self.epoch_limit = np.Inf
        else:
        Severity: Major
        Found in ptp/workers/online_trainer.py and 2 other locations - About 1 hr to fix
        ptp/workers/offline_trainer.py on lines 74..79
        ptp/workers/online_trainer.py on lines 77..82

        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 44.

        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 3 locations. Consider refactoring.
        Open

                if self.early_stop_validations <= 0:
            log_str += "   II: Termination based on Early Stopping is disabled\n"
            # Set to infinity.
            self.early_stop_validations = np.Inf
        else:
        Severity: Major
        Found in ptp/workers/online_trainer.py and 2 other locations - About 1 hr to fix
        ptp/workers/offline_trainer.py on lines 74..79
        ptp/workers/online_trainer.py on lines 86..91

        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 44.

        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.app_state.episode % self.app_state.args.logging_interval == 0:
                        self.logger.info(self.training_stat_col.export_to_string())
        Severity: Major
        Found in ptp/workers/online_trainer.py and 1 other location - About 1 hr to fix
        ptp/workers/offline_trainer.py on lines 243..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 43.

        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.curric_done = self.training.task.curriculum_learning_update_params(
                    0 if self.app_state.episode < 0 else self.app_state.episode,
                    self.app_state.epoch)
        Severity: Major
        Found in ptp/workers/online_trainer.py and 1 other location - About 1 hr to fix
        ptp/workers/offline_trainer.py on lines 171..173

        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 42.

        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

                            try:
                        # if present - clip gradients to a range (-gradient_clipping, gradient_clipping)
                        val = self.config_training['gradient_clipping']
                        torch.nn.utils.clip_grad_value_(self.pipeline.parameters(), val)
                    except KeyError:
        Severity: Major
        Found in ptp/workers/online_trainer.py and 1 other location - About 1 hr to fix
        ptp/workers/offline_trainer.py on lines 202..206

        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 41.

        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

                try:
            '''
            Main training and validation loop.
            '''
            # Reset the counters.
        Severity: Minor
        Found in ptp/workers/online_trainer.py and 1 other location - About 50 mins to fix
        ptp/workers/offline_trainer.py on lines 149..349

        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

        Similar blocks of code found in 4 locations. Consider refactoring.
        Open

                        if self.app_state.epoch+1 >= self.epoch_limit: # = np.Inf when inactive.
                    training_status = "Not converged: Epoch Limit reached"
                    # "Finish" the training.
                    raise TerminationCondition(training_status)
        Severity: Major
        Found in ptp/workers/online_trainer.py and 3 other locations - About 40 mins to fix
        ptp/workers/offline_trainer.py on lines 255..259
        ptp/workers/offline_trainer.py on lines 309..312
        ptp/workers/online_trainer.py on lines 283..287

        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 34.

        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 4 locations. Consider refactoring.
        Open

                            if self.app_state.episode+1 >= self.episode_limit:
                        # If we reach this condition, then it is possible that the model didn't converge correctly
                        # but it currently might get better since last validation.
                        training_status = "Not converged: Episode Limit reached"
                        raise TerminationCondition(training_status)
        Severity: Major
        Found in ptp/workers/online_trainer.py and 3 other locations - About 40 mins to fix
        ptp/workers/offline_trainer.py on lines 255..259
        ptp/workers/offline_trainer.py on lines 309..312
        ptp/workers/online_trainer.py on lines 305..308

        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 34.

        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