vak.transforms.functional

Functional forms of input transforms.

Functions

add_channel(input[, channel_dim])	Add a "channel" dimension to a tensor.
pad_to_window(arr, window_size[, padval, ...])	Pad a 1d or 2d array so that it can be reshaped into consecutive windows of specified size.
standardize_spect(spect, mean_freqs, ...)	standardize spectrogram by subtracting off mean and dividing by standard deviation.
to_floattensor(arr)	convert Numpy array to torch.FloatTensor.
to_longtensor(arr)	convert Numpy array to torch.LongTensor.
view_as_window_batch(arr, window_size)	return view of a 1d or 2d array as a batch of non-overlapping windows