api-transforms.rst

audtorch.transforms

The transforms can be provided to :py:class:`audtorch.datasets` as an argument and work on the data before it will be returned.

Note

All of the transforms work currently only with :py:obj:`numpy.array` as inputs, not :py:obj:`torch.Tensor`.

.. automodule:: audtorch.transforms

Compose

.. autoclass:: Compose
    :members:

Crop

.. autoclass:: Crop
    :members:

RandomCrop

.. autoclass:: RandomCrop
    :members:

Pad

.. autoclass:: Pad
    :members:

RandomPad

.. autoclass:: RandomPad
    :members:

Replicate

.. autoclass:: Replicate
    :members:

RandomReplicate

.. autoclass:: RandomReplicate
    :members:

Expand

.. autoclass:: Expand
    :members:

Downmix

.. autoclass:: Downmix
    :members:

Upmix

.. autoclass:: Upmix
    :members:

Remix

.. autoclass:: Remix
    :members:

Normalize

.. autoclass:: Normalize
    :members:

Resample

.. autoclass:: Resample
    :members:

Spectrogram

.. autoclass:: Spectrogram
    :members:

LogSpectrogram

.. autoclass:: LogSpectrogram
    :members:

RandomAdditiveMix

.. autoclass:: RandomAdditiveMix
    :members: