TimeShift#

class myoverse.transforms.TimeShift(max_shift=100, p=0.5, fill='zero', **kwargs)[source]#

Randomly shift signal in time.

Parameters:

max_shift (int | float) – Maximum shift amount. If float, interpreted as fraction of length.
p (float) – Probability of applying shift.
fill (str) – How to fill shifted regions: ‘zero’, ‘wrap’, ‘edge’.

Examples

>>> x = torch.randn(64, 2048, device='cuda', names=('channel', 'time'))
>>> shift = TimeShift(max_shift=100, p=0.5)
>>> y = shift(x)

Methods

`__init__`([max_shift, p, fill])
`_apply`(x)	Apply the transform.