Postprocessing after the whole step-by-step autoregressive decoding

Postprocessing (save output values) after each timestep t

Train AudioSet tagging model. Args: dataset_dir: str workspace: str data_type: 'balanced_train' | 'full_train' window_si

Handle forward for each training case (distributed, single gpu, etc...) :param batch: :param batch_idx: :return:

Ensures that segments without voice in the waveform remain no longer than a threshold determined by the VAD parameters in params.py. :p

Args: windows: (num_windows*B, window_size, window_size, C) window_size (int): Window size H (int): Height of image

# from librosa 0.6 Compute the sum-square envelope of a window function at a given hop length. This is used to estimate modulation effect

Zero out the parameters of a module and return it.

Get ndarray for a given key.

Load waveform and target of an audio clip. Args: meta: { 'hdf5_path': str, 'index_in_hdf5': int}