Module ktrain.vision.learner

Classes

class ImageClassLearner (model, train_data=None, val_data=None, batch_size=32, eval_batch_size=32, workers=1, use_multiprocessing=False)

Main class used to tune and train Keras models for image classification.
Main parameters are:

model (Model): A compiled instance of keras.engine.training.Model
train_data (Iterator): a Iterator instance for training set
val_data (Iterator):   A Iterator instance for validation set

Expand source code

class ImageClassLearner(GenLearner):
    """
    ```
    Main class used to tune and train Keras models for image classification.
    Main parameters are:

    model (Model): A compiled instance of keras.engine.training.Model
    train_data (Iterator): a Iterator instance for training set
    val_data (Iterator):   A Iterator instance for validation set
    ```
    """

    def __init__(
        self,
        model,
        train_data=None,
        val_data=None,
        batch_size=U.DEFAULT_BS,
        eval_batch_size=U.DEFAULT_BS,
        workers=1,
        use_multiprocessing=False,
    ):
        super().__init__(
            model,
            train_data=train_data,
            val_data=val_data,
            batch_size=batch_size,
            eval_batch_size=eval_batch_size,
            workers=workers,
            use_multiprocessing=use_multiprocessing,
        )
        return

    def view_top_losses(self, n=4, preproc=None, val_data=None):
        """
        ```
        Views observations with top losses in validation set.
        Args:
         n(int or tuple): a range to select in form of int or tuple
                          e.g., n=8 is treated as n=(0,8)
         preproc (Preprocessor): A TextPreprocessor or ImagePreprocessor.
                                 For some data like text data, a preprocessor
                                 is required to undo the pre-processing
                                 to correctly view raw data.
          val_data:  optional val_data to use instead of self.val_data
        Returns:
            list of n tuples where first element is either
            filepath or id of validation example and second element
            is loss.
        ```
        """
        val = self._check_val(val_data)

        # check validation data and arguments
        if val_data is not None:
            val = val_data
        else:
            val = self.val_data
        if val is None:
            raise Exception(
                "val_data must be supplied to get_learner or view_top_losses"
            )

        # get top losses and associated data
        tups = self.top_losses(n=n, val_data=val, preproc=preproc)

        # get multilabel status and class names
        classes = preproc.get_classes() if preproc is not None else None

        # iterate through losses
        for tup in tups:
            # get data
            idx = tup[0]
            loss = tup[1]
            truth = tup[2]
            pred = tup[3]

            # Image Classification
            if type(val).__name__ in ["DirectoryIterator", "DataFrameIterator"]:
                fpath = val.filepaths[tup[0]]
                fp = os.path.join(
                    os.path.basename(os.path.dirname(fpath)), os.path.basename(fpath)
                )
                plt.figure()
                plt.title(
                    "%s | loss:%s | true:%s | pred:%s)"
                    % (fp, round(loss, 2), truth, pred)
                )
                show_image(fpath)
            elif type(val).__name__ in ["NumpyArrayIterator"]:
                obs = val.x[idx]
                # if preproc is not None: obs = preproc.undo(obs)
                plt.figure()
                plt.title(
                    "id:%s | loss:%s | true:%s | pred:%s)"
                    % (idx, round(loss, 2), truth, pred)
                )
                plt.imshow(np.squeeze(obs))
                # everything else including text classification
            else:
                raise Exception(
                    "ImageClassLearner.view_top_losses only supports "
                    + "DirectoryIterators, DataFrameIterators, and NumpyArrayIterators"
                )
        return

Ancestors

• GenLearner
• Learner
• abc.ABC

Methods

def view_top_losses(self, n=4, preproc=None, val_data=None)

Views observations with top losses in validation set.
Args:
 n(int or tuple): a range to select in form of int or tuple
                  e.g., n=8 is treated as n=(0,8)
 preproc (Preprocessor): A TextPreprocessor or ImagePreprocessor.
                         For some data like text data, a preprocessor
                         is required to undo the pre-processing
                         to correctly view raw data.
  val_data:  optional val_data to use instead of self.val_data
Returns:
    list of n tuples where first element is either
    filepath or id of validation example and second element
    is loss.

Expand source code

def view_top_losses(self, n=4, preproc=None, val_data=None):
    """
    ```
    Views observations with top losses in validation set.
    Args:
     n(int or tuple): a range to select in form of int or tuple
                      e.g., n=8 is treated as n=(0,8)
     preproc (Preprocessor): A TextPreprocessor or ImagePreprocessor.
                             For some data like text data, a preprocessor
                             is required to undo the pre-processing
                             to correctly view raw data.
      val_data:  optional val_data to use instead of self.val_data
    Returns:
        list of n tuples where first element is either
        filepath or id of validation example and second element
        is loss.
    ```
    """
    val = self._check_val(val_data)

    # check validation data and arguments
    if val_data is not None:
        val = val_data
    else:
        val = self.val_data
    if val is None:
        raise Exception(
            "val_data must be supplied to get_learner or view_top_losses"
        )

    # get top losses and associated data
    tups = self.top_losses(n=n, val_data=val, preproc=preproc)

    # get multilabel status and class names
    classes = preproc.get_classes() if preproc is not None else None

    # iterate through losses
    for tup in tups:
        # get data
        idx = tup[0]
        loss = tup[1]
        truth = tup[2]
        pred = tup[3]

        # Image Classification
        if type(val).__name__ in ["DirectoryIterator", "DataFrameIterator"]:
            fpath = val.filepaths[tup[0]]
            fp = os.path.join(
                os.path.basename(os.path.dirname(fpath)), os.path.basename(fpath)
            )
            plt.figure()
            plt.title(
                "%s | loss:%s | true:%s | pred:%s)"
                % (fp, round(loss, 2), truth, pred)
            )
            show_image(fpath)
        elif type(val).__name__ in ["NumpyArrayIterator"]:
            obs = val.x[idx]
            # if preproc is not None: obs = preproc.undo(obs)
            plt.figure()
            plt.title(
                "id:%s | loss:%s | true:%s | pred:%s)"
                % (idx, round(loss, 2), truth, pred)
            )
            plt.imshow(np.squeeze(obs))
            # everything else including text classification
        else:
            raise Exception(
                "ImageClassLearner.view_top_losses only supports "
                + "DirectoryIterators, DataFrameIterators, and NumpyArrayIterators"
            )
    return

Inherited members

• GenLearner:
  • autofit
  • evaluate
  • fit
  • fit_onecycle
  • freeze
  • get_weight_decay
  • layer_output
  • load_model
  • lr_estimate
  • lr_find
  • lr_plot
  • plot
  • predict
  • print_layers
  • reset_weights
  • save_model
  • set_model
  • set_weight_decay
  • top_losses
  • unfreeze
  • validate