nam.data

Functions and classes for working with audio data with NAM

Functions

apply_joint_dataset_hooks(dataset_train, ...)
get_joint_dataset_hooks(hook_configs)
init_dataset(config, split)
normalize_joint_dataset_output(*args, **kwargs)
np_to_wav(x, filename[, rate, sampwidth, scale])
register_dataset_initializer(name, constructor)	If you have other data set types, you can register their initializer by name using this.
tensor_to_wav(x, *args, **kwargs)
wav_to_np(filename[, rate, require_match, ...])	param filename: Where to load from
wav_to_tensor(*args[, info])

Classes

AbstractDataset()
ConcatDataset(datasets[, flatten])
Dataset(x, y, nx, ny[, start, stop, ...])	Take a pair of matched audio files and serve input + output pairs.
JointDatasetHook()
NormalizeJointDatasetOutput(level_rms_dbfs)
Split(value)	An enumeration.
WavInfo(sampwidth, rate)	param sampwidth: Sample width, in bytes. May be None if not known.

Exceptions

AudioShapeMismatchError(shape_expected, ...)	Exception where the shape (number of samples, number of channels) of two audio files don't match but were supposed to.
ConcatDatasetValidationError	Error raised when a ConcatDataset fails validation
DataError	Parent class for all special exceptions raised by NAM data sets
DatasetModelHandshakeError	Raised if a handshake fails from dataset to model
JointDatasetValidationError
StartError
StartStopError	Exceptions related to invalid start and stop arguments
StopError
XYError	Exceptions related to invalid x and y provided for data sets