Print confusion matrix as text to terminal

This is a function to process the data, i.e. convert dataframe to numpy array :param X: :param normalise: :return:

Input: args: arguments object from argparse Returns: config: configuration dictionary

Accuracy when allowing for correct class being in the top k predictions. Arguments: predictions: (N_samples, k) array of top class indice

For an array of label predictions and the respective true labels, shows confusion matrix, accuracy, recall, precision etc: Input:

Input: dirs: a list of directories to create, in case these directories are not found Returns: exit_code: 0 if success, -1 if

Returns a string of a report for given metric arrays (array length equals the number of classes). Called internally by `analyze_class

Get average recall and precision, using class frequencies as weights, optionally excluding specified classes

Returns the squared L2 norm of output layer of given model

Wraps a list of metric functions, i.e precison or recall, by ingoring predictions under the specified threshold for a specific class.

Run an evaluation on the validation set while logging metrics, and handle outcome

A simple accuracy calculator, which can ignore labels specified in a list

Print things to stdout on one line, refreshing it dynamically

r"""Inputs of forward function Args: x: the sequence fed to the positional encoder model (required). Shape: x:

r"""Inputs of forward function Args: x: the sequence fed to the positional encoder model (required). Shape: x:

r"""Inputs of forward function Args: x: the sequence fed to the positional encoder model (required). Shape: x:

Make a histogram with the distribution of classes with respect to precision and recall

Returns an array showing for a specified class of interest, how precision and recall change as a function of the score threshold (parameter).

Returns precision for the specified class index

Convert a float measuring time difference in seconds into a tuple of (hours, minutes, seconds)

Returns recall for the specified class index

Show how accuracy improves when considering the event of the correct label being among the top k predictions as a successful prediction

Print the runtime of the decorated function