Initialize a layer

convert dataloader to iterator :param iterable: :return:

Returns mixed inputs, pairs of targets, and lambda

Initialize a GRU layer.

Reads description file and returns classification based SED labels and regression based DOA labels :param _desc_file: metadata description f

Loads DCASE output format csv file and returns it in dictionary format :param _output_format_file: DCASE output format CSV :return: _out

Calculate precisions for each true class for a single sample. Args: scores: np.array of (num_classes,) giving the individual classifier sco

Calculate label-weighted label-ranking average precision. Arguments: truth: np.array of (num_samples, num_classes) giving boolean ground-tr

Computes SED metrics for one second segments :param pred: predicted matrix of dimension [nb_frames, nb_classes], with 1 when sound event is

Converts the seld labels predicted in classification format to dcase output format. :param _feat_cls: feature or data generator class instan

Reads description file in csv format. Outputs, the dcase format results in dictionary, and additionally writes it to the _output_file :p

Interpolate the x to have equal time steps as targets Input: x: (batch_size, time_steps, class_num) Output: out: (batch

Loads DCASE output format csv file and returns it in dictionary format :param _output_format_file: DCASE output format CSV :return: _out

Writes DCASE output format csv file, given output format dictionary :param _output_format_file: :param _output_format_dict: :return: